lab3_impl

CMakeLists.txt

@@ -39,6 +39,7 @@ INCLUDE_DIRECTORIES(
     include/cminusfc
     include/common
     include/lightir
+    include/codegen
     ${LLVM_INCLUDE_DIRS}
 )
 

include/codegen/ASMInstruction.hpp

+#pragma once
+
+#include <cassert>
+#include <string>
+
+struct ASMInstruction {
+    enum InstType { Instruction, Atrribute, Label, Comment } type;
+    std::string content;
+
+    explicit ASMInstruction(std::string s, InstType ty = Instruction)
+        : type(ty), content(s) {}
+
+    std::string format() const {
+        switch (type) {
+        case ASMInstruction::Instruction:
+        case ASMInstruction::Atrribute:
+            return "\t" + content + "\n";
+        case ASMInstruction::Label:
+            return content + ":\n";
+        case ASMInstruction::Comment:
+            return "# " + content + "\n";
+        }
+        assert(false && "unreachable");
+    }
+};

include/codegen/CodeGen.hpp

+#pragma once
+
+#include "ASMInstruction.hpp"
+#include "Module.hpp"
+#include "Register.hpp"
+
+class CodeGen {
+  public:
+    explicit CodeGen(Module *module) : m(module) {}
+
+    std::string print() const;
+
+    void run();
+
+    template <class... Args> void append_inst(Args... arg) {
+        output.emplace_back(arg...);
+    }
+
+    void
+    append_inst(const char *inst, std::initializer_list<std::string> args,
+                ASMInstruction::InstType ty = ASMInstruction::Instruction) {
+        auto content = std::string(inst) + " ";
+        for (const auto &arg : args) {
+            content += arg + ", ";
+        }
+        content.pop_back();
+        content.pop_back();
+        output.emplace_back(content, ty);
+    }
+
+  private:
+    void allocate();
+    void copy_stmt(); // for phi copy
+
+    // 向寄存器中装载数据
+    void load_to_greg(Value *, const Reg &);
+    void load_to_freg(Value *, const FReg &);
+    void load_from_stack_to_greg(Value *, const Reg &);
+
+    // 向寄存器中加载立即数
+    void load_large_int32(int32_t, const Reg &);
+    void load_large_int64(int64_t, const Reg &);
+    void load_float_imm(float, const FReg &);
+
+    // 将寄存器中的数据保存回栈上
+    void store_from_greg(Value *, const Reg &);
+    void store_from_freg(Value *, const FReg &);
+
+    void gen_prologue();
+    void gen_ret();
+    void gen_br();
+    void gen_binary();
+    void gen_float_binary();
+    void gen_alloca();
+    void gen_load();
+    void gen_store();
+    void gen_icmp();
+    void gen_fcmp();
+    void gen_zext();
+    void gen_call();
+    void gen_gep();
+    void gen_sitofp();
+    void gen_fptosi();
+    void gen_epilogue();
+
+    static std::string label_name(BasicBlock *bb) {
+        return "." + bb->get_parent()->get_name() + "_" + bb->get_name();
+    }
+
+    static std::string func_exit_label_name(Function *func) {
+        return func->get_name() + "_exit";
+    }
+
+    static std::string fcmp_label_name(BasicBlock *bb, unsigned cnt) {
+        return label_name(bb) + "_fcmp_" + std::to_string(cnt);
+    }
+
+    struct {
+        /* 随着ir遍历设置 */
+        Function *func{nullptr};    // 当前函数
+        BasicBlock *bb{nullptr};    // 当前基本块
+        Instruction *inst{nullptr}; // 当前指令
+        /* 在allocate()中设置 */
+        unsigned frame_size{0}; // 当前函数的栈帧大小
+        std::unordered_map<Value *, int> offset_map{}; // 指针相对 fp 的偏移
+        unsigned fcmp_cnt{0}; // fcmp 的计数器, 用于创建 fcmp 需要的 label
+
+        void clear() {
+            func = nullptr;
+            bb = nullptr;
+            inst = nullptr;
+            frame_size = 0;
+            fcmp_cnt = 0;
+            offset_map.clear();
+        }
+
+    } context;
+
+    Module *m;
+    std::list<ASMInstruction> output;
+};

include/codegen/CodeGenUtil.hpp

+#pragma once
+
+#include <stdexcept>
+
+/* 关于位宽 */
+#define IMM_12_MAX 0x7FF
+#define IMM_12_MIN -0x800
+
+#define LOW_12_MASK 0x00000FFF
+#define LOW_20_MASK 0x000FFFFF
+#define LOW_32_MASK 0xFFFFFFFF
+
+inline unsigned ALIGN(unsigned x, unsigned alignment) {
+    return ((x + (alignment - 1)) & ~(alignment - 1));
+}
+
+inline bool IS_IMM_12(int x) { return x <= IMM_12_MAX and x >= IMM_12_MIN; }
+
+/* 栈帧相关 */
+#define PROLOGUE_OFFSET_BASE 16 // $ra $fp
+#define PROLOGUE_ALIGN 16
+
+/* 龙芯指令 */
+// Arithmetic
+#define ADD "add"
+#define SUB "sub"
+#define MUL "mul"
+#define DIV "div"
+
+#define ADDI "addi"
+
+#define FADD "fadd"
+#define FSUB "fsub"
+#define FMUL "fmul"
+#define FDIV "fdiv"
+
+#define ORI "ori"
+
+#define LU12I_W "lu12i.w"
+#define LU32I_D "lu32i.d"
+#define LU52I_D "lu52i.d"
+
+// Data transfer (greg <-> freg)
+#define GR2FR "movgr2fr"
+#define FR2GR "movfr2gr"
+
+// Memory access
+#define LOAD "ld"
+#define STORE "st"
+#define FLOAD "fld"
+#define FSTORE "fst"
+
+#define BYTE ".b"
+#define HALF_WORD ".h"
+#define WORD ".w"
+#define DOUBLE ".d"
+
+#define SINGLE ".s" // float
+#define LONG ".l"
+
+// ASM syntax sugar
+#define LOAD_ADDR "la.local"
+
+// errors
+class not_implemented_error : public std::logic_error {
+  public:
+    explicit not_implemented_error(std::string &&err_msg = "")
+        : std::logic_error(err_msg){};
+};
+
+class unreachable_error : public std::logic_error {
+  public:
+    explicit unreachable_error(std::string &&err_msg = "")
+        : std::logic_error(err_msg){};
+};

include/codegen/Register.hpp

+#pragma once
+
+#include <cassert>
+#include <string>
+
+/* General-purpose Register Convention:
+ * Name         Alias       Meaning
+ * $r0          $zero       constant 0
+ * $r1          $ra         return address
+ * $r2          $tp         thread pointer
+ * $r3          $sp         stack pointer
+ * $r4 - $r5    $a0 - $a1   argument, return value
+ * $r6 - $r11   $a2 - $a7   argument
+ * $r12 - $r20  $t0 - $t8   temporary
+ * $r21                     saved
+ * $r22         $fp / $s9   frame pointer
+ * $r23 - $r31  $s0 - $s8   static
+ *
+ * Floating-point Register Convention
+ * Name	        Alias	    Meaning
+ * $f0-$f1      $fa0-$fa1   argument/return value
+ * $f2-$f7      $fa2-$fa7   argument
+ * $f8-$f23     $ft0-$ft15  temporary
+ * $f24-$f31    $fs0-$fs7   static
+ */
+
+struct Reg {
+    unsigned id;
+
+    explicit Reg(unsigned i) : id(i) { assert(i <= 31); }
+    bool operator==(const Reg &other) { return id == other.id; }
+
+    std::string print() const;
+
+    static Reg zero() { return Reg(0); }
+    static Reg ra() { return Reg(1); }
+    static Reg sp() { return Reg(3); }
+    static Reg fp() { return Reg(22); }
+    static Reg a(unsigned i) {
+        assert(0 <= i and i <= 7);
+        return Reg(i + 4);
+    }
+    static Reg t(unsigned i) {
+        assert(0 <= i and i <= 8);
+        return Reg(i + 12);
+    }
+    static Reg s(unsigned i) {
+        assert(0 <= i and i <= 9);
+        if (i == 9)
+            return Reg(22);
+        else
+            return Reg(i + 23);
+    }
+};
+
+struct FReg {
+    unsigned id;
+
+    explicit FReg(unsigned i) : id(i) { assert(i <= 31); }
+    bool operator==(const FReg &other) { return id == other.id; }
+
+    std::string print() const;
+
+    static FReg fa(unsigned i) {
+        assert(0 <= i and i <= 7);
+        return FReg(i);
+    }
+    static FReg ft(unsigned i) {
+        assert(0 <= i and i <= 15);
+        return FReg(i + 8);
+    }
+    static FReg fs(unsigned i) {
+        assert(0 <= i and i <= 7);
+        return FReg(i + 24);
+    }
+};
+
+struct CFReg {
+    unsigned id;
+
+    explicit CFReg(unsigned i) : id(i) { assert(i <= 7); }
+    bool operator==(const CFReg &other) { return id == other.id; }
+
+    std::string print() const { return "$fcc" + std::to_string(id); }
+};

src/CMakeLists.txt

@@ -4,3 +4,4 @@ add_subdirectory(logging)
 add_subdirectory(cminusfc)
 add_subdirectory(lightir)
 add_subdirectory(io)
+add_subdirectory(codegen)
\ No newline at end of file

src/cminusfc/CMakeLists.txt

@@ -9,6 +9,7 @@ target_link_libraries(
     IR_lib
     common
     syntax
+    codegen
     stdc++fs
 )
 

src/cminusfc/main.cpp

 #include "Module.hpp"
 #include "ast.hpp"
 #include "cminusf_builder.hpp"
+#include "CodeGen.hpp"
 
 #include <filesystem>
 #include <fstream>
@@ -16,6 +17,7 @@ struct Config {
     std::filesystem::path output_file;
 
     bool emitast{false};
+    bool emitasm{false};
     bool emitllvm{false};
 
     Config(int argc, char **argv) : argc(argc), argv(argv) {
@@ -55,9 +57,13 @@ int main(int argc, char **argv) {
             output_stream << "; ModuleID = 'cminus'\n";
             output_stream << "source_filename = " << abs_path << "\n\n";
             output_stream << m->print();
+        } else if (config.emitasm) {
+            CodeGen codegen(m.get());
+            codegen.run();
+            output_stream << codegen.print();
         }
 
-        // TODO: lab3 lab4 (IR optimization or codegen)
+        // TODO: lab4 (IR optimization or codegen)
     }
 
     return 0;
@@ -77,6 +83,8 @@ void Config::parse_cmd_line() {
             }
         } else if (argv[i] == "-emit-ast"s) {
             emitast = true;
+        } else if (argv[i] == "-S"s) {
+            emitasm = true;
         } else if (argv[i] == "-emit-llvm"s) {
             emitllvm = true;
         } else {
@@ -105,7 +113,7 @@ void Config::check() {
 
 void Config::print_help() const {
     std::cout << "Usage: " << exe_name
-              << " [-h|--help] [-o <target-file>] [-mem2reg] [-emit-llvm] [-S] "
+              << " [-h|--help] [-o <target-file>] [-emit-ast] [-emit-llvm] [-S] "
                  "<input-file>"
               << std::endl;
     exit(0);

src/codegen/CMakeLists.txt

+add_library(
+    codegen STATIC
+    CodeGen.cpp
+    Register.cpp
+)
+
+target_link_libraries(codegen common IR_lib)

src/codegen/CodeGen.cpp

+#include "CodeGen.hpp"
+
+#include "CodeGenUtil.hpp"
+
+void CodeGen::allocate() {
+    // 备份 $ra $fp
+    unsigned offset = PROLOGUE_OFFSET_BASE;
+
+    // 为每个参数分配栈空间
+    for (auto &arg : context.func->get_args()) {
+        auto size = arg.get_type()->get_size();
+        offset = offset + size;
+        context.offset_map[&arg] = -static_cast<int>(offset);
+    }
+
+    // 为指令结果分配栈空间
+    for (auto &bb : context.func->get_basic_blocks()) {
+        for (auto &instr : bb.get_instructions()) {
+            // 每个非 void 的定值都分配栈空间
+            if (not instr.is_void()) {
+                auto size = instr.get_type()->get_size();
+                offset = offset + size;
+                context.offset_map[&instr] = -static_cast<int>(offset);
+            }
+            // alloca 的副作用：分配额外空间
+            if (instr.is_alloca()) {
+                auto *alloca_inst = static_cast<AllocaInst *>(&instr);
+                auto alloc_size = alloca_inst->get_alloca_type()->get_size();
+                offset += alloc_size;
+            }
+        }
+    }
+
+    // 分配栈空间，需要是 16 的整数倍
+    context.frame_size = ALIGN(offset, PROLOGUE_ALIGN);
+}
+
+void CodeGen::copy_stmt() {
+    for (auto &succ : context.bb->get_succ_basic_blocks()) {
+        for (auto &inst : succ->get_instructions()) {
+            if (inst.is_phi()) {
+                // 遍历后继块中 phi 的定值 bb
+                for (unsigned i = 1; i < inst.get_operands().size(); i += 2) {
+                    // phi 的定值 bb 是当前翻译块
+                    if (inst.get_operand(i) == context.bb) {
+                        auto *lvalue = inst.get_operand(i - 1);
+                        if (lvalue->get_type()->is_float_type()) {
+                            load_to_freg(lvalue, FReg::fa(0));
+                            store_from_freg(&inst, FReg::fa(0));
+                        } else {
+                            load_to_greg(lvalue, Reg::a(0));
+                            store_from_greg(&inst, Reg::a(0));
+                        }
+                        break;
+                    }
+                    // 如果没有找到当前翻译块，说明是 undef，无事可做
+                }
+            } else {
+                break;
+            }
+        }
+    }
+}
+
+void CodeGen::load_to_greg(Value *val, const Reg &reg) {
+    assert(val->get_type()->is_integer_type() ||
+           val->get_type()->is_pointer_type());
+
+    if (auto *constant = dynamic_cast<ConstantInt *>(val)) {
+        int32_t val = constant->get_value();
+        if (IS_IMM_12(val)) {
+            append_inst(ADDI WORD, {reg.print(), "$zero", std::to_string(val)});
+        } else {
+            load_large_int32(val, reg);
+        }
+    } else if (auto *global = dynamic_cast<GlobalVariable *>(val)) {
+        append_inst(LOAD_ADDR, {reg.print(), global->get_name()});
+    } else {
+        load_from_stack_to_greg(val, reg);
+    }
+}
+
+void CodeGen::load_large_int32(int32_t val, const Reg &reg) {
+    int32_t high_20 = val >> 12; // si20
+    uint32_t low_12 = val & LOW_12_MASK;
+    append_inst(LU12I_W, {reg.print(), std::to_string(high_20)});
+    append_inst(ORI, {reg.print(), reg.print(), std::to_string(low_12)});
+}
+
+void CodeGen::load_large_int64(int64_t val, const Reg &reg) {
+    auto low_32 = static_cast<int32_t>(val & LOW_32_MASK);
+    load_large_int32(low_32, reg);
+
+    auto high_32 = static_cast<int32_t>(val >> 32);
+    int32_t high_32_low_20 = (high_32 << 12) >> 12; // si20
+    int32_t high_32_high_12 = high_32 >> 20;        // si12
+    append_inst(LU32I_D, {reg.print(), std::to_string(high_32_low_20)});
+    append_inst(LU52I_D,
+                {reg.print(), reg.print(), std::to_string(high_32_high_12)});
+}
+
+void CodeGen::load_from_stack_to_greg(Value *val, const Reg &reg) {
+    auto offset = context.offset_map.at(val);
+    auto offset_str = std::to_string(offset);
+    auto *type = val->get_type();
+    if (IS_IMM_12(offset)) {
+        if (type->is_int1_type()) {
+            append_inst(LOAD BYTE, {reg.print(), "$fp", offset_str});
+        } else if (type->is_int32_type()) {
+            append_inst(LOAD WORD, {reg.print(), "$fp", offset_str});
+        } else { // Pointer
+            append_inst(LOAD DOUBLE, {reg.print(), "$fp", offset_str});
+        }
+    } else {
+        load_large_int64(offset, reg);
+        append_inst(ADD DOUBLE, {reg.print(), "$fp", reg.print()});
+        if (type->is_int1_type()) {
+            append_inst(LOAD BYTE, {reg.print(), reg.print(), "0"});
+        } else if (type->is_int32_type()) {
+            append_inst(LOAD WORD, {reg.print(), reg.print(), "0"});
+        } else { // Pointer
+            append_inst(LOAD DOUBLE, {reg.print(), reg.print(), "0"});
+        }
+    }
+}
+
+void CodeGen::store_from_greg(Value *val, const Reg &reg) {
+    auto offset = context.offset_map.at(val);
+    auto offset_str = std::to_string(offset);
+    auto *type = val->get_type();
+    if (IS_IMM_12(offset)) {
+        if (type->is_int1_type()) {
+            append_inst(STORE BYTE, {reg.print(), "$fp", offset_str});
+        } else if (type->is_int32_type()) {
+            append_inst(STORE WORD, {reg.print(), "$fp", offset_str});
+        } else { // Pointer
+            append_inst(STORE DOUBLE, {reg.print(), "$fp", offset_str});
+        }
+    } else {
+        auto addr = Reg::t(8);
+        load_large_int64(offset, addr);
+        append_inst(ADD DOUBLE, {addr.print(), "$fp", addr.print()});
+        if (type->is_int1_type()) {
+            append_inst(STORE BYTE, {reg.print(), addr.print(), "0"});
+        } else if (type->is_int32_type()) {
+            append_inst(STORE WORD, {reg.print(), addr.print(), "0"});
+        } else { // Pointer
+            append_inst(STORE DOUBLE, {reg.print(), addr.print(), "0"});
+        }
+    }
+}
+
+void CodeGen::load_to_freg(Value *val, const FReg &freg) {
+    assert(val->get_type()->is_float_type());
+    if (auto *constant = dynamic_cast<ConstantFP *>(val)) {
+        float val = constant->get_value();
+        load_float_imm(val, freg);
+    } else {
+        auto offset = context.offset_map.at(val);
+        auto offset_str = std::to_string(offset);
+        if (IS_IMM_12(offset)) {
+            append_inst(FLOAD SINGLE, {freg.print(), "$fp", offset_str});
+        } else {
+            auto addr = Reg::t(8);
+            load_large_int64(offset, addr);
+            append_inst(ADD DOUBLE, {addr.print(), "$fp", addr.print()});
+            append_inst(FLOAD SINGLE, {freg.print(), addr.print(), "0"});
+        }
+    }
+}
+
+void CodeGen::load_float_imm(float val, const FReg &r) {
+    int32_t bytes = *reinterpret_cast<int32_t *>(&val);
+    load_large_int32(bytes, Reg::t(8));
+    append_inst(GR2FR WORD, {r.print(), Reg::t(8).print()});
+}
+
+void CodeGen::store_from_freg(Value *val, const FReg &r) {
+    auto offset = context.offset_map.at(val);
+    if (IS_IMM_12(offset)) {
+        auto offset_str = std::to_string(offset);
+        append_inst(FSTORE SINGLE, {r.print(), "$fp", offset_str});
+    } else {
+        auto addr = Reg::t(8);
+        load_large_int64(offset, addr);
+        append_inst(ADD DOUBLE, {addr.print(), "$fp", addr.print()});
+        append_inst(FSTORE SINGLE, {r.print(), addr.print(), "0"});
+    }
+}
+
+void CodeGen::gen_prologue() {
+    if (IS_IMM_12(-static_cast<int>(context.frame_size))) {
+        append_inst("st.d $ra, $sp, -8");
+        append_inst("st.d $fp, $sp, -16");
+        append_inst("addi.d $fp, $sp, 0");
+        append_inst("addi.d $sp, $sp, " +
+                    std::to_string(-static_cast<int>(context.frame_size)));
+    } else {
+        load_large_int64(context.frame_size, Reg::t(0));
+        append_inst("st.d $ra, $sp, -8");
+        append_inst("st.d $fp, $sp, -16");
+        append_inst("sub.d $sp, $sp, $t0");
+        append_inst("add.d $fp, $sp, $t0");
+    }
+
+    int garg_cnt = 0;
+    int farg_cnt = 0;
+    for (auto &arg : context.func->get_args()) {
+        if (arg.get_type()->is_float_type()) {
+            store_from_freg(&arg, FReg::fa(farg_cnt++));
+        } else { // int or pointer
+            store_from_greg(&arg, Reg::a(garg_cnt++));
+        }
+    }
+}
+
+// void CodeGen::gen_epilogue() {
+//     // TODO 根据你的理解设定函数的 epilogue
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+
+// void CodeGen::gen_ret() {
+//     // TODO 函数返回，思考如何处理返回值、寄存器备份，如何返回调用者地址
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_br() {
+//     auto *branchInst = static_cast<BranchInst *>(context.inst);
+//     if (branchInst->is_cond_br()) {
+//         // TODO 补全条件跳转的情况
+//         throw not_implemented_error{__FUNCTION__};
+//     } else {
+//         auto *branchbb = static_cast<BasicBlock *>(branchInst->get_operand(0));
+//         append_inst("b " + label_name(branchbb));
+//     }
+// }
+
+// void CodeGen::gen_binary() {
+//     load_to_greg(context.inst->get_operand(0), Reg::t(0));
+//     load_to_greg(context.inst->get_operand(1), Reg::t(1));
+//     switch (context.inst->get_instr_type()) {
+//     case Instruction::add:
+//         output.emplace_back("add.w $t2, $t0, $t1");
+//         break;
+//     case Instruction::sub:
+//         output.emplace_back("sub.w $t2, $t0, $t1");
+//         break;
+//     case Instruction::mul:
+//         output.emplace_back("mul.w $t2, $t0, $t1");
+//         break;
+//     case Instruction::sdiv:
+//         output.emplace_back("div.w $t2, $t0, $t1");
+//         break;
+//     default:
+//         assert(false);
+//     }
+//     store_from_greg(context.inst, Reg::t(2));
+// }
+
+// void CodeGen::gen_float_binary() {
+//     // TODO 浮点类型的二元指令
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_alloca() {
+//     /* 我们已经为 alloca 的内容分配空间，在此我们还需保存 alloca
+//      * 指令自身产生的定值，即指向 alloca 空间起始地址的指针
+//      */
+//     // TODO 将 alloca 出空间的起始地址保存在栈帧上
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_load() {
+//     auto *ptr = context.inst->get_operand(0);
+//     auto *type = context.inst->get_type();
+//     load_to_greg(ptr, Reg::t(0));
+
+//     if (type->is_float_type()) {
+//         append_inst("fld.s $ft0, $t0, 0");
+//         store_from_freg(context.inst, FReg::ft(0));
+//     } else {
+//         // TODO load 整数类型的数据
+//         throw not_implemented_error{__FUNCTION__};
+//     }
+// }
+
+// void CodeGen::gen_store() {
+//     // TODO 翻译 store 指令
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_icmp() {
+//     // TODO 处理各种整数比较的情况
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_fcmp() {
+//     // TODO 处理各种浮点数比较的情况
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_zext() {
+//     // TODO 将窄位宽的整数数据进行零扩展
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_call() {
+//     // TODO 函数调用，注意我们只需要通过寄存器传递参数，即不需考虑栈上传参的情况
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// /*
+//  * %op = getelementptr [10 x i32], [10 x i32]* %op, i32 0, i32 %op
+//  * %op = getelementptr        i32,        i32* %op, i32 %op
+//  *
+//  * Memory layout
+//  *       -            ^
+//  * +-----------+      | Smaller address
+//  * |  arg ptr  |---+  |
+//  * +-----------+   |  |
+//  * |           |   |  |
+//  * +-----------+   /  |
+//  * |           |<--   |
+//  * |           |   \  |
+//  * |           |   |  |
+//  * |   Array   |   |  |
+//  * |           |   |  |
+//  * |           |   |  |
+//  * |           |   |  |
+//  * +-----------+   |  |
+//  * |  Pointer  |---+  |
+//  * +-----------+      |
+//  * |           |      |
+//  * +-----------+      |
+//  * |           |      |
+//  * +-----------+      |
+//  * |           |      |
+//  * +-----------+      | Larger address
+//  *       +
+//  */
+// void CodeGen::gen_gep() {
+//     // TODO 计算内存地址
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_sitofp() {
+//     // TODO 整数转向浮点数
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::gen_fptosi() {
+//     // TODO 浮点数转向整数，注意向下取整(round to zero)
+//     throw not_implemented_error{__FUNCTION__};
+// }
+
+// void CodeGen::run() {
+//     // 确保每个函数中基本块的名字都被设置好
+//     m->set_print_name();
+
+//     /* 使用 GNU 伪指令为全局变量分配空间
+//      * 你可以使用 `la.local` 指令将标签 (全局变量) 的地址载入寄存器中, 比如
+//      * 要将 `a` 的地址载入 $t0, 只需要 `la.local $t0, a`
+//      */
+//     if (!m->get_global_variable().empty()) {
+//         append_inst("Global variables", ASMInstruction::Comment);
+//         /* 虽然下面两条伪指令可以简化为一条 `.bss` 伪指令, 但是我们还是选择使用
+//          * `.section` 将全局变量放到可执行文件的 BSS 段, 原因如下:
+//          * - 尽可能对齐交叉编译器 loongarch64-unknown-linux-gnu-gcc 的行为
+//          * - 支持更旧版本的 GNU 汇编器, 因为 `.bss` 伪指令是应该相对较新的指令,
+//          *   GNU 汇编器在 2023 年 2 月的 2.37 版本才将其引入
+//          */
+//         append_inst(".text", ASMInstruction::Atrribute);
+//         append_inst(".section", {".bss", "\"aw\"", "@nobits"},
+//                     ASMInstruction::Atrribute);
+//         for (auto &global : m->get_global_variable()) {
+//             auto size =
+//                 global.get_type()->get_pointer_element_type()->get_size();
+//             append_inst(".globl", {global.get_name()},
+//                         ASMInstruction::Atrribute);
+//             append_inst(".type", {global.get_name(), "@object"},
+//                         ASMInstruction::Atrribute);
+//             append_inst(".size", {global.get_name(), std::to_string(size)},
+//                         ASMInstruction::Atrribute);
+//             append_inst(global.get_name(), ASMInstruction::Label);
+//             append_inst(".space", {std::to_string(size)},
+//                         ASMInstruction::Atrribute);
+//         }
+//     }
+
+//     // 函数代码段
+//     output.emplace_back(".text", ASMInstruction::Atrribute);
+//     for (auto &func : m->get_functions()) {
+//         if (not func.is_declaration()) {
+//             // 更新 context
+//             context.clear();
+//             context.func = &func;
+
+//             // 函数信息
+//             append_inst(".globl", {func.get_name()}, ASMInstruction::Atrribute);
+//             append_inst(".type", {func.get_name(), "@function"},
+//                         ASMInstruction::Atrribute);
+//             append_inst(func.get_name(), ASMInstruction::Label);
+
+//             // 分配函数栈帧
+//             allocate();
+//             // 生成 prologue
+//             gen_prologue();
+
+//             for (auto &bb : func.get_basic_blocks()) {
+//                 context.bb = &bb;
+//                 append_inst(label_name(context.bb), ASMInstruction::Label);
+//                 for (auto &instr : bb.get_instructions()) {
+//                     // For debug
+//                     append_inst(instr.print(), ASMInstruction::Comment);
+//                     context.inst = &instr; // 更新 context
+//                     switch (instr.get_instr_type()) {
+//                     case Instruction::ret:
+//                         gen_ret();
+//                         break;
+//                     case Instruction::br:
+//                         copy_stmt();
+//                         gen_br();
+//                         break;
+//                     case Instruction::add:
+//                     case Instruction::sub:
+//                     case Instruction::mul:
+//                     case Instruction::sdiv:
+//                         gen_binary();
+//                         break;
+//                     case Instruction::fadd:
+//                     case Instruction::fsub:
+//                     case Instruction::fmul:
+//                     case Instruction::fdiv:
+//                         gen_float_binary();
+//                         break;
+//                     case Instruction::alloca:
+//                         /* 对于 alloca 指令，我们已经为 alloca
+//                          * 的内容分配空间，在此我们还需保存 alloca
+//                          * 指令自身产生的定值，即指向 alloca 空间起始地址的指针
+//                          */
+//                         gen_alloca();
+//                         break;
+//                     case Instruction::load:
+//                         gen_load();
+//                         break;
+//                     case Instruction::store:
+//                         gen_store();
+//                         break;
+//                     case Instruction::ge:
+//                     case Instruction::gt:
+//                     case Instruction::le:
+//                     case Instruction::lt:
+//                     case Instruction::eq:
+//                     case Instruction::ne:
+//                         gen_icmp();
+//                         break;
+//                     case Instruction::fge:
+//                     case Instruction::fgt:
+//                     case Instruction::fle:
+//                     case Instruction::flt:
+//                     case Instruction::feq:
+//                     case Instruction::fne:
+//                         gen_fcmp();
+//                         break;
+//                     case Instruction::phi:
+//                         /* for phi, just convert to a series of
+//                          * copy-stmts */
+//                         /* we can collect all phi and deal them at
+//                          * the end */
+//                         break;
+//                     case Instruction::call:
+//                         gen_call();
+//                         break;
+//                     case Instruction::getelementptr:
+//                         gen_gep();
+//                         break;
+//                     case Instruction::zext:
+//                         gen_zext();
+//                         break;
+//                     case Instruction::fptosi:
+//                         gen_fptosi();
+//                         break;
+//                     case Instruction::sitofp:
+//                         gen_sitofp();
+//                         break;
+//                     }
+//                 }
+//             }
+//             // 生成 epilogue
+//             gen_epilogue();
+//         }
+//     }
+// }
+
+// std::string CodeGen::print() const {
+//     std::string result;
+//     for (const auto &inst : output) {
+//         result += inst.format();
+//     }
+//     return result;
+// }
+void CodeGen::gen_epilogue() {
+    append_inst(func_exit_label_name(context.func), ASMInstruction::Label);
+
+    if (IS_IMM_12(static_cast<int>(context.frame_size))) {
+        append_inst("addi.d $sp, $sp, " + std::to_string(context.frame_size));
+        append_inst("ld.d $ra, $sp, -8");
+        append_inst("ld.d $fp, $sp, -16");
+    } else {
+        load_large_int64(context.frame_size, Reg::t(0));
+        append_inst("add.d $sp, $sp, $t0");
+        append_inst("ld.d $ra, $sp, -8");
+        append_inst("ld.d $fp, $sp, -16");
+    }
+
+    output.emplace_back("jr $ra");
+}
+
+void CodeGen::gen_ret() {
+    if (not context.func->get_return_type()->is_void_type()) {
+        // 需要把返回值放到 $a0 中
+        auto *ret_val = context.inst->get_operand(0);
+        if (ret_val->get_type()->is_float_type()) {
+            load_to_freg(ret_val, FReg::fa(0));
+        } else {
+            load_to_greg(ret_val, Reg::a(0));
+        }
+    } else {
+        append_inst("addi.w $a0, $zero, 0");
+    }
+    append_inst("b " + func_exit_label_name(context.func));
+}
+
+void CodeGen::gen_br() {
+    auto *branchInst = static_cast<BranchInst *>(context.inst);
+    if (branchInst->is_cond_br()) {
+        auto *cond = branchInst->get_operand(0);
+        auto *truebb = static_cast<BasicBlock *>(branchInst->get_operand(1));
+        auto *falsebb = static_cast<BasicBlock *>(branchInst->get_operand(2));
+        load_to_greg(cond, Reg::t(0));
+        // FIXME: Acutually this instruction is not necessary since
+        // load will sign extend the value
+        append_inst("bstrpick.d $t0, $t0, 0, 0");
+        append_inst("bne $t0, $zero, " + label_name(truebb));
+        append_inst("b " + label_name(falsebb));
+    } else {
+        auto *branchbb = static_cast<BasicBlock *>(branchInst->get_operand(0));
+        append_inst("b " + label_name(branchbb));
+    }
+}
+
+void CodeGen::gen_binary() {
+    load_to_greg(context.inst->get_operand(0), Reg::t(0));
+    load_to_greg(context.inst->get_operand(1), Reg::t(1));
+    switch (context.inst->get_instr_type()) {
+    case Instruction::add:
+        output.emplace_back("add.w $t2, $t0, $t1");
+        break;
+    case Instruction::sub:
+        output.emplace_back("sub.w $t2, $t0, $t1");
+        break;
+    case Instruction::mul:
+        output.emplace_back("mul.w $t2, $t0, $t1");
+        break;
+    case Instruction::sdiv:
+        output.emplace_back("div.w $t2, $t0, $t1");
+        break;
+    default:
+        assert(false);
+    }
+    store_from_greg(context.inst, Reg::t(2));
+}
+
+void CodeGen::gen_float_binary() {
+    load_to_freg(context.inst->get_operand(0), FReg::ft(0));
+    load_to_freg(context.inst->get_operand(1), FReg::ft(1));
+    switch (context.inst->get_instr_type()) {
+    case Instruction::fadd:
+        output.emplace_back("fadd.s $ft2, $ft0, $ft1");
+        break;
+    case Instruction::fsub:
+        output.emplace_back("fsub.s $ft2, $ft0, $ft1");
+        break;
+    case Instruction::fmul:
+        output.emplace_back("fmul.s $ft2, $ft0, $ft1");
+        break;
+    case Instruction::fdiv:
+        output.emplace_back("fdiv.s $ft2, $ft0, $ft1");
+        break;
+    default:
+        assert(false);
+    }
+    store_from_freg(context.inst, FReg::ft(2));
+}
+
+void CodeGen::gen_alloca() {
+    /* Current strategy: The allocated space is always after the
+     * pointer in the stack, so we can just add the size of the
+     * pointer to the offset.
+     * TODO: consider better strategy */
+    auto size = static_cast<int>(
+        context.inst->get_type()->get_pointer_element_type()->get_size());
+    auto offset = context.offset_map.at(context.inst);
+    auto position = offset - size;
+    if (IS_IMM_12(position))
+        append_inst("addi.d $t0, $fp, " + std::to_string(position));
+    else {
+        load_large_int32(position, Reg::t(0));
+        append_inst("add.d $t0, $fp, $t0");
+    }
+    store_from_greg(context.inst, Reg::t(0));
+    // append_inst("st.d $t0, $fp, " + std::to_string(offset));
+}
+
+void CodeGen::gen_load() {
+    auto *ptr = context.inst->get_operand(0);
+    auto *type = context.inst->get_type();
+    load_to_greg(ptr, Reg::t(0));
+
+    if (type->is_float_type()) {
+        append_inst("fld.s $ft0, $t0, 0");
+        store_from_freg(context.inst, FReg::ft(0));
+    } else {
+        if (type->is_int1_type()) {
+            append_inst("ld.b $t0, $t0, 0");
+        } else if (type->is_int32_type()) {
+            append_inst("ld.w $t0, $t0, 0");
+        } else { // Pointer
+            append_inst("ld.d $t0, $t0, 0");
+        }
+
+        store_from_greg(context.inst, Reg::t(0));
+    }
+}
+
+void CodeGen::gen_store() {
+    auto *ptr = context.inst->get_operand(1);
+    load_to_greg(ptr, Reg::t(0));
+
+    auto *v = context.inst->get_operand(0);
+
+    if (v->get_type()->is_float_type()) {
+        load_to_freg(v, FReg::ft(0));
+        append_inst("fst.s $ft0, $t0, 0");
+    } else {
+        load_to_greg(v, Reg::t(1));
+        if (v->get_type()->is_int1_type()) {
+            append_inst("st.b $t1, $t0, 0");
+        } else if (v->get_type()->is_int32_type()) {
+            append_inst("st.w $t1, $t0, 0");
+        } else { // Pointer
+            append_inst("st.d $t1, $t0, 0");
+        }
+    }
+}
+
+void CodeGen::gen_icmp() {
+    // CAUTION: only slt is available
+    load_to_greg(context.inst->get_operand(0), Reg::t(0));
+    load_to_greg(context.inst->get_operand(1), Reg::t(1));
+    switch (context.inst->get_instr_type()) {
+    case Instruction::ge:
+        // t0 >= t1 <=> !(t0 < t1) <=> (t0 < t1) xor 1
+        append_inst("slt $t2, $t0, $t1");
+        append_inst("xori $t2, $t2, 1");
+        break;
+    case Instruction::gt:
+        // t0 > t1 <=> t1 < t0
+        append_inst("slt $t2, $t1, $t0");
+        break;
+    case Instruction::le:
+        // t0 <= t1 <=> !(t1 < t0) <=> (t1 < t0) xor 1
+        append_inst("slt $t2, $t1, $t0");
+        append_inst("xori $t2, $t2, 1");
+        break;
+    case Instruction::lt:
+        // t0 < t1
+        append_inst("slt $t2, $t0, $t1");
+        break;
+    case Instruction::eq:
+        // t0 == t1 <=> (t0 xor t1) == 0 <=> (unsigned)(t0 xor t1) <
+        // 1
+        append_inst("xor $t2, $t0, $t1");
+        append_inst("sltui $t2, $t2, 1");
+        break;
+    case Instruction::ne:
+        // t0 != t1 <=> (t0 xor t1) != 0 <=> (unsigned)(t0 xor t1) >
+        // 0
+        append_inst("xor $t2, $t0, $t1");
+        append_inst("sltu $t2, $zero, $t2");
+        break;
+
+    default:
+        assert(false);
+    }
+    store_from_greg(context.inst, Reg::t(2));
+}
+
+void CodeGen::gen_fcmp() {
+    load_to_freg(context.inst->get_operand(0), FReg::ft(0));
+    load_to_freg(context.inst->get_operand(1), FReg::ft(1));
+    switch (context.inst->get_instr_type()) {
+    case Instruction::fge:
+        append_inst("fcmp.sle.s $fcc0, $ft1, $ft0");
+        break;
+    case Instruction::fgt:
+        append_inst("fcmp.slt.s $fcc0, $ft1, $ft0");
+        break;
+    case Instruction::fle:
+        append_inst("fcmp.sle.s $fcc0, $ft0, $ft1");
+        break;
+    case Instruction::flt:
+        append_inst("fcmp.slt.s $fcc0, $ft0, $ft1");
+        break;
+    case Instruction::feq:
+        append_inst("fcmp.seq.s $fcc0, $ft0, $ft1");
+        break;
+    case Instruction::fne:
+        append_inst("fcmp.sne.s $fcc0, $ft0, $ft1");
+        break;
+
+    default:
+        assert(false);
+    }
+    // Branch on FP Condition
+    append_inst("addi.w $t0, $zero, 1");
+    auto label = fcmp_label_name(context.bb, context.fcmp_cnt++);
+    append_inst("bcnez $fcc0, " + label);
+    append_inst("addi.w $t0, $zero, 0");
+    append_inst(label, ASMInstruction::Label);
+    store_from_greg(context.inst, Reg::t(0));
+}
+
+void CodeGen::gen_zext() {
+    auto *source = context.inst->get_operand(0);
+    auto bits = static_cast<IntegerType *>(source->get_type())->get_num_bits();
+    load_to_greg(source, Reg::t(0));
+    // FIXME: Acutually this instruction is not necessary
+    append_inst("bstrpick.w $t0, $t0, " + std::to_string(bits - 1) + ", 0");
+    store_from_greg(context.inst, Reg::t(0));
+}
+
+void CodeGen::gen_call() {
+    auto operands = context.inst->get_operands();
+    auto *func = static_cast<Function *>(operands[0]);
+    // 把参数放进 a 或 fa 寄存器
+    // 注意我们只允许至多八个整形和至多八个浮点参数
+    auto n = operands.size();
+    int garg_cnt = 0;
+    int farg_cnt = 0;
+    for (size_t i = 1; i < n; ++i) {
+        auto *arg = operands[i];
+        if (arg->get_type()->is_float_type()) {
+            load_to_freg(arg, FReg::fa(farg_cnt++));
+        } else { // int or pointer
+            load_to_greg(arg, Reg::a(garg_cnt++));
+        }
+    }
+
+    append_inst("bl " + func->get_name());
+    if (not func->get_return_type()->is_void_type()) {
+        if (func->get_return_type()->is_float_type()) {
+            store_from_freg(context.inst, FReg::fa(0));
+        } else {
+            store_from_greg(context.inst, Reg::a(0));
+        }
+    }
+}
+
+/*
+ * %op = getelementptr [10 x i32], [10 x i32]* %op, i32 0, i32 %op
+ * %op = getelementptr        i32,        i32* %op, i32 %op
+ *
+ * Memory layout
+ *       -            ^
+ * +-----------+      | Smaller address
+ * |  arg ptr  |---+  |
+ * +-----------+   |  |
+ * |           |   |  |
+ * +-----------+   /  |
+ * |           |<--   |
+ * |           |   \  |
+ * |           |   |  |
+ * |   Array   |   |  |
+ * |           |   |  |
+ * |           |   |  |
+ * |           |   |  |
+ * +-----------+   |  |
+ * |  Pointer  |---+  |
+ * +-----------+      |
+ * |           |      |
+ * +-----------+      |
+ * |           |      |
+ * +-----------+      |
+ * |           |      |
+ * +-----------+      | Larger address
+ *       +
+ */
+void CodeGen::gen_gep() {
+    auto *ptr = context.inst->get_operand(0);
+    load_to_greg(ptr, Reg::t(0));
+    // 这时 $t0 里应该是数组的起始地址
+    auto *type = ptr->get_type()->get_pointer_element_type(); // [10 x i32]
+    auto operands = context.inst->get_operands();
+    auto n = operands.size();
+    for (size_t i = 1; i < n; ++i) {
+        // $t1 里是各偏移量
+        load_to_greg(operands[i], Reg::t(1));
+
+        auto size = type->get_size();
+        // $t0 += index * size
+        if (IS_IMM_12(static_cast<int>(size))) {
+            append_inst("addi.d $t2, $zero, " + std::to_string(size));
+        } else {
+            load_large_int64(static_cast<int64_t>(size), Reg::t(2));
+        }
+
+        append_inst("mul.d $t1, $t1, $t2");
+        append_inst("add.d $t0, $t0, $t1");
+
+        if (type->is_array_type()) {
+            type = type->get_array_element_type(); // 1st it: type
+                                                   // <- i32
+        } else if (type->is_pointer_type()) {
+            type = type->get_pointer_element_type();
+        }
+    }
+
+    store_from_greg(context.inst, Reg::t(0));
+}
+
+void CodeGen::gen_sitofp() {
+    auto *operand = context.inst->get_operand(0);
+    load_to_greg(operand, Reg::t(0));
+    append_inst("movgr2fr.w $ft0, $t0");
+    append_inst("ffint.s.w $ft0, $ft0");
+    store_from_freg(context.inst, FReg::ft(0));
+}
+
+void CodeGen::gen_fptosi() {
+    // The LLVM IR fptosi uses round-to-zero mode
+    auto *operand = context.inst->get_operand(0);
+    load_to_freg(operand, FReg::ft(0));
+    append_inst("ftintrz.w.s $ft0, $ft0");
+    append_inst("movfr2gr.s $t0, $ft0");
+    store_from_greg(context.inst, Reg::t(0));
+}
+
+void CodeGen::run() {
+    // 确保每个函数中基本块的名字都被设置好
+    m->set_print_name();
+
+    /* 使用 GNU 伪指令为全局变量分配空间
+     * 你可以使用 `la.local` 指令将标签 (全局变量) 的地址载入寄存器中, 比如
+     * 要将 `a` 的地址载入 $t0, 只需要 `la.local $t0, a`
+     */
+    if (!m->get_global_variable().empty()) {
+        append_inst("Global variables", ASMInstruction::Comment);
+        /* 虽然下面两条伪指令可以简化为一条 `.bss` 伪指令, 但是我们还是选择使用
+         * `.section` 将全局变量放到可执行文件的 BSS 段, 原因如下:
+         * - 尽可能对齐交叉编译器 loongarch64-unknown-linux-gnu-gcc 的行为
+         * - 支持更旧版本的 GNU 汇编器, 因为 `.bss` 伪指令是应该相对较新的指令,
+         *   GNU 汇编器在 2023 年 2 月的 2.37 版本才将其引入
+         */
+        append_inst(".text", ASMInstruction::Atrribute);
+        append_inst(".section", {".bss", "\"aw\"", "@nobits"},
+                    ASMInstruction::Atrribute);
+        for (auto &global : m->get_global_variable()) {
+            auto size =
+                global.get_type()->get_pointer_element_type()->get_size();
+            append_inst(".globl", {global.get_name()},
+                        ASMInstruction::Atrribute);
+            append_inst(".type", {global.get_name(), "@object"},
+                        ASMInstruction::Atrribute);
+            append_inst(".size", {global.get_name(), std::to_string(size)},
+                        ASMInstruction::Atrribute);
+            append_inst(global.get_name(), ASMInstruction::Label);
+            append_inst(".space", {std::to_string(size)},
+                        ASMInstruction::Atrribute);
+        }
+    }
+
+    // 函数代码段
+    output.emplace_back(".text", ASMInstruction::Atrribute);
+    for (auto &func : m->get_functions()) {
+        if (not func.is_declaration()) {
+            // 更新 context
+            context.clear();
+            context.func = &func;
+
+            // 函数信息
+            append_inst(".globl", {func.get_name()}, ASMInstruction::Atrribute);
+            append_inst(".type", {func.get_name(), "@function"},
+                        ASMInstruction::Atrribute);
+            append_inst(func.get_name(), ASMInstruction::Label);
+
+            // 分配函数栈帧
+            allocate();
+            // 生成 prologue
+            gen_prologue();
+
+            for (auto &bb : func.get_basic_blocks()) {
+                context.bb = &bb;
+                append_inst(label_name(context.bb), ASMInstruction::Label);
+                for (auto &instr : bb.get_instructions()) {
+                    // For debug
+                    append_inst(instr.print(), ASMInstruction::Comment);
+                    context.inst = &instr; // 更新 context
+                    switch (instr.get_instr_type()) {
+                    case Instruction::ret:
+                        gen_ret();
+                        break;
+                    case Instruction::br:
+                        copy_stmt();
+                        gen_br();
+                        break;
+                    case Instruction::add:
+                    case Instruction::sub:
+                    case Instruction::mul:
+                    case Instruction::sdiv:
+                        gen_binary();
+                        break;
+                    case Instruction::fadd:
+                    case Instruction::fsub:
+                    case Instruction::fmul:
+                    case Instruction::fdiv:
+                        gen_float_binary();
+                        break;
+                    case Instruction::alloca:
+                        /* 对于 alloca 指令，我们已经为 alloca
+                         * 的内容分配空间，在此我们还需保存 alloca
+                         * 指令自身产生的定值，即指向 alloca 空间起始地址的指针
+                         */
+                        gen_alloca();
+                        break;
+                    case Instruction::load:
+                        gen_load();
+                        break;
+                    case Instruction::store:
+                        gen_store();
+                        break;
+                    case Instruction::ge:
+                    case Instruction::gt:
+                    case Instruction::le:
+                    case Instruction::lt:
+                    case Instruction::eq:
+                    case Instruction::ne:
+                        gen_icmp();
+                        break;
+                    case Instruction::fge:
+                    case Instruction::fgt:
+                    case Instruction::fle:
+                    case Instruction::flt:
+                    case Instruction::feq:
+                    case Instruction::fne:
+                        gen_fcmp();
+                        break;
+                    case Instruction::phi:
+                        /* for phi, just convert to a series of
+                         * copy-stmts */
+                        /* we can collect all phi and deal them at
+                         * the end */
+                        break;
+                    case Instruction::call:
+                        gen_call();
+                        break;
+                    case Instruction::getelementptr:
+                        gen_gep();
+                        break;
+                    case Instruction::zext:
+                        gen_zext();
+                        break;
+                    case Instruction::fptosi:
+                        gen_fptosi();
+                        break;
+                    case Instruction::sitofp:
+                        gen_sitofp();
+                        break;
+                    }
+                }
+            }
+            // 生成 epilogue
+            gen_epilogue();
+        }
+    }
+}
+
+std::string CodeGen::print() const {
+    std::string result;
+    for (const auto &inst : output) {
+        result += inst.format();
+    }
+    return result;
+}
\ No newline at end of file

src/codegen/Register.cpp

+#include "Register.hpp"
+#include <string>
+
+std::string Reg::print() const {
+    if (id == 0) {
+        return "$zero";
+    }
+    if (id == 1) {
+        return "$ra";
+    }
+    if (id == 2) {
+        return "$tp";
+    }
+    if (id == 3) {
+        return "$sp";
+    }
+    if (4 <= id and id <= 11) {
+        return "$a" + std::to_string(id - 4);
+    }
+    if (12 <= id and id <= 20) {
+        return "$t" + std::to_string(id - 12);
+    }
+    if (id == 22) {
+        return "$fp";
+    }
+    assert(false);
+}
+
+std::string FReg::print() const {
+
+    if (0 <= id and id <= 7) {
+        return "$fa" + std::to_string(id);
+    }
+    if (8 <= id and id <= 23) {
+        return "$ft" + std::to_string(id - 8);
+    }
+    if (24 <= id and id <= 31) {
+        return "$fs" + std::to_string(id - 24);
+    }
+    assert(false);
+}

tests/3-codegen/autogen/cleanup.sh

+#!/bin/bash
+rm -rf output log.txt

tests/3-codegen/autogen/eval_lab3.sh

+#!/bin/bash
+
+project_dir=$(realpath ../../../)
+io_dir=$(realpath "$project_dir"/src/io)
+output_dir=output
+suffix=cminus
+
+LOG=log.txt
+
+usage() {
+	cat <<JIANMU
+Usage: $0 [path-to-testcases] [type]
+path-to-testcases: './testcases' or '../../testcases_general' or 'self made cases'
+type: 'debug' or 'test', debug will output .ll file
+JIANMU
+	exit 0
+}
+
+check_return_value() {
+	rv=$1
+	expected_rv=$2
+	fail_msg=$3
+	detail=$4
+	if [ "$rv" -eq "$expected_rv" ]; then
+		return 0
+	else
+		printf "\033[1;31m%s: \033[0m%s\n" "$fail_msg" "$detail"
+		return 1
+	fi
+}
+
+# check arguments
+[ $# -lt 2 ] && usage
+if [ "$2" == "debug" ]; then
+	debug_mode=true
+elif [ "$2" == "test" ]; then
+	debug_mode=false
+else
+	usage
+fi
+
+test_dir=$1
+testcases=$(ls "$test_dir"/*."$suffix" | sort -V)
+check_return_value $? 0 "PATH" "unable to access to '$test_dir'" || exit 1
+
+# hide stderr in the script
+# exec 2>/dev/null
+
+mkdir -p $output_dir
+
+truncate -s 0 $LOG
+
+if [ $debug_mode = false ]; then
+	exec 3>/dev/null 4>&1 5>&2 1>&3 2>&3
+else
+	exec 3>&1
+fi
+
+if [ $debug_mode = false ]; then
+	exec 1>&4 2>&5
+fi
+
+echo "[info] Start testing, using testcase dir: $test_dir"
+# asm
+for case in $testcases; do
+	echo "==========$case==========" >>$LOG
+	case_base_name=$(basename -s .$suffix "$case")
+	std_out_file=$test_dir/$case_base_name.out
+	in_file=$test_dir/$case_base_name.in
+	asm_file=$output_dir/$case_base_name.s
+	exe_file=$output_dir/$case_base_name
+	out_file=$output_dir/$case_base_name.out
+	ll_file=$output_dir/$case_base_name.ll
+
+	echo -n "$case_base_name..."
+	# if debug mode on, generate .ll also
+	if [ $debug_mode = true ]; then
+		bash -c "${project_dir}/build/cminusfc -emit-llvm $case -o $ll_file" >>$LOG 2>&1
+	fi
+	# cminusfc compile to .s
+	bash -c "${project_dir}/build/cminusfc -S $case -o $asm_file" >>$LOG 2>&1
+	check_return_value $? 0 "CE" "cminusfc compiler error" || continue
+
+	# gcc compile asm to executable
+	loongarch64-unknown-linux-gnu-gcc -static \
+		"$asm_file" "$io_dir"/io.c -o "$exe_file" \
+		>>$LOG
+	check_return_value $? 0 "CE" "gcc compiler error" || continue
+
+	# qemu run
+	if [ -e "$in_file" ]; then
+		exec_cmd="qemu-loongarch64 $exe_file >$out_file <$in_file"
+	else
+		exec_cmd="qemu-loongarch64 $exe_file >$out_file"
+	fi
+	bash -c "$exec_cmd"
+	ret=$?
+	# remove trailing null byte in the end line
+	sed -i "\$s/\x00*$//" "$out_file"
+	# append return value
+	echo $ret >>"$out_file"
+
+	# compare output
+	diff --strip-trailing-cr "$std_out_file" "$out_file" -y >>$LOG
+	check_return_value $? 0 "WA" "output differ, check $std_out_file and $out_file" || continue
+
+	# ok
+	printf "\033[1;32mOK\033[0m\n"
+done

tests/3-codegen/autogen/log.txt

+==========../../testcases_general//1-return.cminus==========
+0								0
+==========../../testcases_general//2-decl_int.cminus==========
+0								0
+==========../../testcases_general//3-decl_float.cminus==========
+0								0
+==========../../testcases_general//4-decl_int_array.cminus==========
+0								0
+==========../../testcases_general//5-decl_float_array.cminus==========
+0								0
+==========../../testcases_general//6-num_add_int.cminus==========
+210								210
+==========../../testcases_general//7-assign_int_var_local.cminus==========
+210								210
+==========../../testcases_general//8-assign_int_array_local.cminus==========
+210								210
+==========../../testcases_general//9-assign_cast.cminus==========
+5								5
+==========../../testcases_general//10-funcall.cminus==========
+0								0
+==========../../testcases_general//11-funcall_chain.cminus==========
+210								210
+==========../../testcases_general//12-funcall_recursion.cminus==========
+0								0
+==========../../testcases_general//13-if_stmt.cminus==========
+0								0
+==========../../testcases_general//14-while_stmt.cminus==========
+0								0
+==========../../testcases_general//15-if_while.cminus==========
+45								45
+==========../../testcases_general//16-if_chain.cminus==========
+3								3
+==========../../testcases_general//17-while_chain.cminus==========
+0								0
+==========../../testcases_general//18-global_var.cminus==========
+10								10
+==========../../testcases_general//19-global_local_var.cminus==========
+20								20
+==========../../testcases_general//20-gcd_array.cminus==========
+18								18
+==========../../testcases_general//21-comment.cminus==========
+0								0

tests/3-codegen/autogen/output/0-io.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/0-io.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  %op0 = call i32 @input()
+  call void @output(i32 %op0)
+  ret i32 0
+}

tests/3-codegen/autogen/output/0-io.out

+1234
+0

tests/3-codegen/autogen/output/0-io.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.main_label_entry:
+# %op0 = call i32 @input()
+	bl input
+	st.w $a0, $fp, -20
+# call void @output(i32 %op0)
+	ld.w $a0, $fp, -20
+	bl output
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/1-return.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/1-return.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  call void @output(i32 111)
+  ret i32 111
+}

tests/3-codegen/autogen/output/1-return.out

+0

tests/3-codegen/autogen/output/1-return.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -16
+.main_label_entry:
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 16
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/10-float.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/10-float.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  %op0 = alloca float
+  %op1 = alloca float
+  %op2 = alloca float
+  store float 0x3ff19999a0000000, float* %op0
+  store float 0x3ff8000000000000, float* %op1
+  store float 0x3ff3333340000000, float* %op2
+  %op3 = load float, float* %op0
+  %op4 = load float, float* %op1
+  %op5 = fmul float %op3, %op4
+  %op6 = load float, float* %op2
+  %op7 = fadd float %op5, %op6
+  call void @outputFloat(float %op7)
+  ret i32 0
+}

tests/3-codegen/autogen/output/10-float.out

+2.850000
+0

tests/3-codegen/autogen/output/10-float.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+.main_label_entry:
+# %op0 = alloca float
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca float
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# %op2 = alloca float
+	addi.d $t0, $fp, -52
+	st.d $t0, $fp, -48
+# store float 0x3ff19999a0000000, float* %op0
+	ld.d $t0, $fp, -24
+	lu12i.w $t8, 260300
+	ori $t8, $t8, 3277
+	movgr2fr.w $ft0, $t8
+	fst.s $ft0, $t0, 0
+# store float 0x3ff8000000000000, float* %op1
+	ld.d $t0, $fp, -36
+	lu12i.w $t8, 261120
+	ori $t8, $t8, 0
+	movgr2fr.w $ft0, $t8
+	fst.s $ft0, $t0, 0
+# store float 0x3ff3333340000000, float* %op2
+	ld.d $t0, $fp, -48
+	lu12i.w $t8, 260505
+	ori $t8, $t8, 2458
+	movgr2fr.w $ft0, $t8
+	fst.s $ft0, $t0, 0
+# %op3 = load float, float* %op0
+	ld.d $t0, $fp, -24
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -56
+# %op4 = load float, float* %op1
+	ld.d $t0, $fp, -36
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -60
+# %op5 = fmul float %op3, %op4
+	fld.s $ft0, $fp, -56
+	fld.s $ft1, $fp, -60
+	fmul.s $ft2, $ft0, $ft1
+	fst.s $ft2, $fp, -64
+# %op6 = load float, float* %op2
+	ld.d $t0, $fp, -48
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -68
+# %op7 = fadd float %op5, %op6
+	fld.s $ft0, $fp, -64
+	fld.s $ft1, $fp, -68
+	fadd.s $ft2, $ft0, $ft1
+	fst.s $ft2, $fp, -72
+# call void @outputFloat(float %op7)
+	fld.s $fa0, $fp, -72
+	bl outputFloat
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/10-funcall.out

+0

tests/3-codegen/autogen/output/10-funcall.s

+	.text
+	.globl test
+	.type test, @function
+test:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+	st.w $a0, $fp, -20
+.test_label_entry:
+# %op1 = alloca i32
+	addi.d $t0, $fp, -32
+	st.d $t0, $fp, -28
+# store i32 %arg0, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# ret void
+	addi.w $a0, $zero, 0
+	b test_exit
+test_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# store i32 10, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# %op1 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -32
+# call void @test(i32 %op1)
+	ld.w $a0, $fp, -32
+	bl test
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/11-floatcall.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/11-floatcall.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define float @mod(float %arg0, float %arg1) {
+label_entry:
+  %op2 = alloca float
+  store float %arg0, float* %op2
+  %op3 = alloca float
+  store float %arg1, float* %op3
+  %op4 = alloca i32
+  %op5 = load float, float* %op2
+  %op6 = load float, float* %op3
+  %op7 = fdiv float %op5, %op6
+  %op8 = fptosi float %op7 to i32
+  store i32 %op8, i32* %op4
+  %op9 = load float, float* %op2
+  %op10 = load i32, i32* %op4
+  %op11 = load float, float* %op3
+  %op12 = sitofp i32 %op10 to float
+  %op13 = fmul float %op12, %op11
+  %op14 = fsub float %op9, %op13
+  ret float %op14
+}
+define i32 @main() {
+label_entry:
+  %op0 = alloca float
+  %op1 = alloca float
+  store float 0x4026666660000000, float* %op0
+  store float 0x40019999a0000000, float* %op1
+  %op2 = load float, float* %op0
+  %op3 = load float, float* %op1
+  %op4 = call float @mod(float %op2, float %op3)
+  call void @outputFloat(float %op4)
+  ret i32 0
+}

tests/3-codegen/autogen/output/11-floatcall.out

+0.200000
+0

tests/3-codegen/autogen/output/11-floatcall.s

+	.text
+	.globl mod
+	.type mod, @function
+mod:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -112
+	fst.s $fa0, $fp, -20
+	fst.s $fa1, $fp, -24
+.mod_label_entry:
+# %op2 = alloca float
+	addi.d $t0, $fp, -36
+	st.d $t0, $fp, -32
+# store float %arg0, float* %op2
+	ld.d $t0, $fp, -32
+	fld.s $ft0, $fp, -20
+	fst.s $ft0, $t0, 0
+# %op3 = alloca float
+	addi.d $t0, $fp, -48
+	st.d $t0, $fp, -44
+# store float %arg1, float* %op3
+	ld.d $t0, $fp, -44
+	fld.s $ft0, $fp, -24
+	fst.s $ft0, $t0, 0
+# %op4 = alloca i32
+	addi.d $t0, $fp, -60
+	st.d $t0, $fp, -56
+# %op5 = load float, float* %op2
+	ld.d $t0, $fp, -32
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -64
+# %op6 = load float, float* %op3
+	ld.d $t0, $fp, -44
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -68
+# %op7 = fdiv float %op5, %op6
+	fld.s $ft0, $fp, -64
+	fld.s $ft1, $fp, -68
+	fdiv.s $ft2, $ft0, $ft1
+	fst.s $ft2, $fp, -72
+# %op8 = fptosi float %op7 to i32
+	fld.s $ft0, $fp, -72
+	ftintrz.w.s $ft0, $ft0
+	movfr2gr.s $t0, $ft0
+	st.w $t0, $fp, -76
+# store i32 %op8, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t1, $fp, -76
+	st.w $t1, $t0, 0
+# %op9 = load float, float* %op2
+	ld.d $t0, $fp, -32
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -80
+# %op10 = load i32, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -84
+# %op11 = load float, float* %op3
+	ld.d $t0, $fp, -44
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -88
+# %op12 = sitofp i32 %op10 to float
+	ld.w $t0, $fp, -84
+	movgr2fr.w $ft0, $t0
+	ffint.s.w $ft0, $ft0
+	fst.s $ft0, $fp, -92
+# %op13 = fmul float %op12, %op11
+	fld.s $ft0, $fp, -92
+	fld.s $ft1, $fp, -88
+	fmul.s $ft2, $ft0, $ft1
+	fst.s $ft2, $fp, -96
+# %op14 = fsub float %op9, %op13
+	fld.s $ft0, $fp, -80
+	fld.s $ft1, $fp, -96
+	fsub.s $ft2, $ft0, $ft1
+	fst.s $ft2, $fp, -100
+# ret float %op14
+	fld.s $fa0, $fp, -100
+	b mod_exit
+mod_exit:
+	addi.d $sp, $sp, 112
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -64
+.main_label_entry:
+# %op0 = alloca float
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca float
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# store float 0x4026666660000000, float* %op0
+	ld.d $t0, $fp, -24
+	lu12i.w $t8, 267059
+	ori $t8, $t8, 819
+	movgr2fr.w $ft0, $t8
+	fst.s $ft0, $t0, 0
+# store float 0x40019999a0000000, float* %op1
+	ld.d $t0, $fp, -36
+	lu12i.w $t8, 262348
+	ori $t8, $t8, 3277
+	movgr2fr.w $ft0, $t8
+	fst.s $ft0, $t0, 0
+# %op2 = load float, float* %op0
+	ld.d $t0, $fp, -24
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -44
+# %op3 = load float, float* %op1
+	ld.d $t0, $fp, -36
+	fld.s $ft0, $t0, 0
+	fst.s $ft0, $fp, -48
+# %op4 = call float @mod(float %op2, float %op3)
+	fld.s $fa0, $fp, -44
+	fld.s $fa1, $fp, -48
+	bl mod
+	fst.s $fa0, $fp, -52
+# call void @outputFloat(float %op4)
+	fld.s $fa0, $fp, -52
+	bl outputFloat
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 64
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/11-funcall_chain.out

+210

tests/3-codegen/autogen/output/11-funcall_chain.s

+	.text
+	.globl addone
+	.type addone, @function
+addone:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -48
+	st.w $a0, $fp, -20
+.addone_label_entry:
+# %op1 = alloca i32
+	addi.d $t0, $fp, -32
+	st.d $t0, $fp, -28
+# store i32 %arg0, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# %op2 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -36
+# %op3 = add i32 %op2, 1
+	ld.w $t0, $fp, -36
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -40
+# ret i32 %op3
+	ld.w $a0, $fp, -40
+	b addone_exit
+addone_exit:
+	addi.d $sp, $sp, 48
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -48
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = call i32 @addone(i32 1230)
+	addi.w $a0, $zero, 1230
+	bl addone
+	st.w $a0, $fp, -32
+# %op2 = call i32 @addone(i32 %op1)
+	ld.w $a0, $fp, -32
+	bl addone
+	st.w $a0, $fp, -36
+# %op3 = call i32 @addone(i32 %op2)
+	ld.w $a0, $fp, -36
+	bl addone
+	st.w $a0, $fp, -40
+# %op4 = call i32 @addone(i32 %op3)
+	ld.w $a0, $fp, -40
+	bl addone
+	st.w $a0, $fp, -44
+# store i32 %op4, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -44
+	st.w $t1, $t0, 0
+# %op5 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -48
+# ret i32 %op5
+	ld.w $a0, $fp, -48
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 48
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/12-funcall_recursion.out

+0

tests/3-codegen/autogen/output/12-funcall_recursion.s

+	.text
+	.globl factorial
+	.type factorial, @function
+factorial:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -64
+	st.w $a0, $fp, -20
+.factorial_label_entry:
+# %op1 = alloca i32
+	addi.d $t0, $fp, -32
+	st.d $t0, $fp, -28
+# store i32 %arg0, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# %op2 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -36
+# %op3 = icmp eq i32 %op2, 0
+	ld.w $t0, $fp, -36
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltui $t2, $t2, 1
+	st.b $t2, $fp, -37
+# %op4 = zext i1 %op3 to i32
+	ld.b $t0, $fp, -37
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -41
+# %op5 = icmp ne i32 %op4, 0
+	ld.w $t0, $fp, -41
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -42
+# br i1 %op5, label %label6, label %label8
+	ld.b $t0, $fp, -42
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .factorial_label6
+	b .factorial_label8
+.factorial_label6:
+# ret i32 1
+	addi.w $a0, $zero, 1
+	b factorial_exit
+.factorial_label7:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b factorial_exit
+.factorial_label8:
+# %op9 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -46
+# %op10 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -50
+# %op11 = sub i32 %op10, 1
+	ld.w $t0, $fp, -50
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -54
+# %op12 = call i32 @factorial(i32 %op11)
+	ld.w $a0, $fp, -54
+	bl factorial
+	st.w $a0, $fp, -58
+# %op13 = mul i32 %op9, %op12
+	ld.w $t0, $fp, -46
+	ld.w $t1, $fp, -58
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -62
+# ret i32 %op13
+	ld.w $a0, $fp, -62
+	b factorial_exit
+factorial_exit:
+	addi.d $sp, $sp, 64
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -48
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = call i32 @factorial(i32 10)
+	addi.w $a0, $zero, 10
+	bl factorial
+	st.w $a0, $fp, -32
+# store i32 %op1, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -32
+	st.w $t1, $t0, 0
+# %op2 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -36
+# ret i32 %op2
+	ld.w $a0, $fp, -36
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 48
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/12-global.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/12-global.cminus"
+
+@seed = global i32 zeroinitializer
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @randomLCG() {
+label_entry:
+  %op0 = load i32, i32* @seed
+  %op1 = mul i32 %op0, 1103515245
+  %op2 = add i32 %op1, 12345
+  store i32 %op2, i32* @seed
+  %op3 = load i32, i32* @seed
+  ret i32 %op3
+}
+define i32 @randBin() {
+label_entry:
+  %op0 = call i32 @randomLCG()
+  %op1 = icmp sgt i32 %op0, 0
+  %op2 = zext i1 %op1 to i32
+  %op3 = icmp ne i32 %op2, 0
+  br i1 %op3, label %label4, label %label6
+label4:                                                ; preds = %label_entry
+  ret i32 1
+label5:
+  ret i32 0
+label6:                                                ; preds = %label_entry
+  ret i32 0
+}
+define i32 @returnToZeroSteps() {
+label_entry:
+  %op0 = alloca i32
+  %op1 = alloca i32
+  store i32 0, i32* %op0
+  store i32 0, i32* %op1
+  br label %label2
+label2:                                                ; preds = %label_entry, %label26
+  %op3 = load i32, i32* %op1
+  %op4 = icmp slt i32 %op3, 20
+  %op5 = zext i1 %op4 to i32
+  %op6 = icmp ne i32 %op5, 0
+  br i1 %op6, label %label7, label %label10
+label7:                                                ; preds = %label2
+  %op8 = call i32 @randBin()
+  %op9 = icmp ne i32 %op8, 0
+  br i1 %op9, label %label11, label %label21
+label10:                                                ; preds = %label2
+  ret i32 20
+label11:                                                ; preds = %label7
+  %op12 = load i32, i32* %op0
+  %op13 = add i32 %op12, 1
+  store i32 %op13, i32* %op0
+  br label %label14
+label14:                                                ; preds = %label11, %label21
+  %op15 = load i32, i32* %op1
+  %op16 = add i32 %op15, 1
+  store i32 %op16, i32* %op1
+  %op17 = load i32, i32* %op0
+  %op18 = icmp eq i32 %op17, 0
+  %op19 = zext i1 %op18 to i32
+  %op20 = icmp ne i32 %op19, 0
+  br i1 %op20, label %label24, label %label26
+label21:                                                ; preds = %label7
+  %op22 = load i32, i32* %op0
+  %op23 = sub i32 %op22, 1
+  store i32 %op23, i32* %op0
+  br label %label14
+label24:                                                ; preds = %label14
+  %op25 = load i32, i32* %op1
+  ret i32 %op25
+label26:                                                ; preds = %label14
+  br label %label2
+}
+define i32 @main() {
+label_entry:
+  %op0 = alloca i32
+  store i32 0, i32* %op0
+  store i32 3407, i32* @seed
+  br label %label1
+label1:                                                ; preds = %label_entry, %label6
+  %op2 = load i32, i32* %op0
+  %op3 = icmp slt i32 %op2, 20
+  %op4 = zext i1 %op3 to i32
+  %op5 = icmp ne i32 %op4, 0
+  br i1 %op5, label %label6, label %label10
+label6:                                                ; preds = %label1
+  %op7 = call i32 @returnToZeroSteps()
+  call void @output(i32 %op7)
+  %op8 = load i32, i32* %op0
+  %op9 = add i32 %op8, 1
+  store i32 %op9, i32* %op0
+  br label %label1
+label10:                                                ; preds = %label1
+  ret i32 0
+}

tests/3-codegen/autogen/output/12-global.out

+4
+2
+2
+4
+8
+2
+2
+2
+2
+2
+6
+2
+10
+8
+4
+2
+20
+2
+2
+8
+0

tests/3-codegen/autogen/output/12-global.s

+# Global variables
+	.text
+	.section .bss, "aw", @nobits
+	.globl seed
+	.type seed, @object
+	.size seed, 4
+seed:
+	.space 4
+	.text
+	.globl randomLCG
+	.type randomLCG, @function
+randomLCG:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.randomLCG_label_entry:
+# %op0 = load i32, i32* @seed
+	la.local $t0, seed
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -20
+# %op1 = mul i32 %op0, 1103515245
+	ld.w $t0, $fp, -20
+	lu12i.w $t1, 269412
+	ori $t1, $t1, 3693
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -24
+# %op2 = add i32 %op1, 12345
+	ld.w $t0, $fp, -24
+	lu12i.w $t1, 3
+	ori $t1, $t1, 57
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -28
+# store i32 %op2, i32* @seed
+	la.local $t0, seed
+	ld.w $t1, $fp, -28
+	st.w $t1, $t0, 0
+# %op3 = load i32, i32* @seed
+	la.local $t0, seed
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -32
+# ret i32 %op3
+	ld.w $a0, $fp, -32
+	b randomLCG_exit
+randomLCG_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl randBin
+	.type randBin, @function
+randBin:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.randBin_label_entry:
+# %op0 = call i32 @randomLCG()
+	bl randomLCG
+	st.w $a0, $fp, -20
+# %op1 = icmp sgt i32 %op0, 0
+	ld.w $t0, $fp, -20
+	addi.w $t1, $zero, 0
+	slt $t2, $t1, $t0
+	st.b $t2, $fp, -21
+# %op2 = zext i1 %op1 to i32
+	ld.b $t0, $fp, -21
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -25
+# %op3 = icmp ne i32 %op2, 0
+	ld.w $t0, $fp, -25
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -26
+# br i1 %op3, label %label4, label %label6
+	ld.b $t0, $fp, -26
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .randBin_label4
+	b .randBin_label6
+.randBin_label4:
+# ret i32 1
+	addi.w $a0, $zero, 1
+	b randBin_exit
+.randBin_label5:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b randBin_exit
+.randBin_label6:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b randBin_exit
+randBin_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl returnToZeroSteps
+	.type returnToZeroSteps, @function
+returnToZeroSteps:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -96
+.returnToZeroSteps_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# store i32 0, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# store i32 0, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# br label %label2
+	b .returnToZeroSteps_label2
+.returnToZeroSteps_label2:
+# %op3 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -44
+# %op4 = icmp slt i32 %op3, 20
+	ld.w $t0, $fp, -44
+	addi.w $t1, $zero, 20
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -45
+# %op5 = zext i1 %op4 to i32
+	ld.b $t0, $fp, -45
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -49
+# %op6 = icmp ne i32 %op5, 0
+	ld.w $t0, $fp, -49
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -50
+# br i1 %op6, label %label7, label %label10
+	ld.b $t0, $fp, -50
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .returnToZeroSteps_label7
+	b .returnToZeroSteps_label10
+.returnToZeroSteps_label7:
+# %op8 = call i32 @randBin()
+	bl randBin
+	st.w $a0, $fp, -54
+# %op9 = icmp ne i32 %op8, 0
+	ld.w $t0, $fp, -54
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -55
+# br i1 %op9, label %label11, label %label21
+	ld.b $t0, $fp, -55
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .returnToZeroSteps_label11
+	b .returnToZeroSteps_label21
+.returnToZeroSteps_label10:
+# ret i32 20
+	addi.w $a0, $zero, 20
+	b returnToZeroSteps_exit
+.returnToZeroSteps_label11:
+# %op12 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -59
+# %op13 = add i32 %op12, 1
+	ld.w $t0, $fp, -59
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -63
+# store i32 %op13, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -63
+	st.w $t1, $t0, 0
+# br label %label14
+	b .returnToZeroSteps_label14
+.returnToZeroSteps_label14:
+# %op15 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -67
+# %op16 = add i32 %op15, 1
+	ld.w $t0, $fp, -67
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -71
+# store i32 %op16, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t1, $fp, -71
+	st.w $t1, $t0, 0
+# %op17 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -75
+# %op18 = icmp eq i32 %op17, 0
+	ld.w $t0, $fp, -75
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltui $t2, $t2, 1
+	st.b $t2, $fp, -76
+# %op19 = zext i1 %op18 to i32
+	ld.b $t0, $fp, -76
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -80
+# %op20 = icmp ne i32 %op19, 0
+	ld.w $t0, $fp, -80
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -81
+# br i1 %op20, label %label24, label %label26
+	ld.b $t0, $fp, -81
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .returnToZeroSteps_label24
+	b .returnToZeroSteps_label26
+.returnToZeroSteps_label21:
+# %op22 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -85
+# %op23 = sub i32 %op22, 1
+	ld.w $t0, $fp, -85
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -89
+# store i32 %op23, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -89
+	st.w $t1, $t0, 0
+# br label %label14
+	b .returnToZeroSteps_label14
+.returnToZeroSteps_label24:
+# %op25 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -93
+# ret i32 %op25
+	ld.w $a0, $fp, -93
+	b returnToZeroSteps_exit
+.returnToZeroSteps_label26:
+# br label %label2
+	b .returnToZeroSteps_label2
+returnToZeroSteps_exit:
+	addi.d $sp, $sp, 96
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -64
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# store i32 0, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# store i32 3407, i32* @seed
+	la.local $t0, seed
+	lu12i.w $t1, 0
+	ori $t1, $t1, 3407
+	st.w $t1, $t0, 0
+# br label %label1
+	b .main_label1
+.main_label1:
+# %op2 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -32
+# %op3 = icmp slt i32 %op2, 20
+	ld.w $t0, $fp, -32
+	addi.w $t1, $zero, 20
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -33
+# %op4 = zext i1 %op3 to i32
+	ld.b $t0, $fp, -33
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -37
+# %op5 = icmp ne i32 %op4, 0
+	ld.w $t0, $fp, -37
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -38
+# br i1 %op5, label %label6, label %label10
+	ld.b $t0, $fp, -38
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label6
+	b .main_label10
+.main_label6:
+# %op7 = call i32 @returnToZeroSteps()
+	bl returnToZeroSteps
+	st.w $a0, $fp, -42
+# call void @output(i32 %op7)
+	ld.w $a0, $fp, -42
+	bl output
+# %op8 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -46
+# %op9 = add i32 %op8, 1
+	ld.w $t0, $fp, -46
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -50
+# store i32 %op9, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -50
+	st.w $t1, $t0, 0
+# br label %label1
+	b .main_label1
+.main_label10:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 64
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/13-complex.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/13-complex.cminus"
+
+@n = global i32 zeroinitializer
+@m = global i32 zeroinitializer
+@w = global [5 x i32] zeroinitializer
+@v = global [5 x i32] zeroinitializer
+@dp = global [66 x i32] zeroinitializer
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @max(i32 %arg0, i32 %arg1) {
+label_entry:
+  %op2 = alloca i32
+  store i32 %arg0, i32* %op2
+  %op3 = alloca i32
+  store i32 %arg1, i32* %op3
+  %op4 = load i32, i32* %op2
+  %op5 = load i32, i32* %op3
+  %op6 = icmp sgt i32 %op4, %op5
+  %op7 = zext i1 %op6 to i32
+  %op8 = icmp ne i32 %op7, 0
+  br i1 %op8, label %label9, label %label12
+label9:                                                ; preds = %label_entry
+  %op10 = load i32, i32* %op2
+  ret i32 %op10
+label11:
+  ret i32 0
+label12:                                                ; preds = %label_entry
+  %op13 = load i32, i32* %op3
+  ret i32 %op13
+}
+define i32 @knapsack(i32 %arg0, i32 %arg1) {
+label_entry:
+  %op2 = alloca i32
+  store i32 %arg0, i32* %op2
+  %op3 = alloca i32
+  store i32 %arg1, i32* %op3
+  %op4 = alloca i32
+  %op5 = load i32, i32* %op3
+  %op6 = icmp sle i32 %op5, 0
+  %op7 = zext i1 %op6 to i32
+  %op8 = icmp ne i32 %op7, 0
+  br i1 %op8, label %label9, label %label10
+label9:                                                ; preds = %label_entry
+  ret i32 0
+label10:                                                ; preds = %label_entry
+  %op11 = load i32, i32* %op2
+  %op12 = icmp eq i32 %op11, 0
+  %op13 = zext i1 %op12 to i32
+  %op14 = icmp ne i32 %op13, 0
+  br i1 %op14, label %label15, label %label16
+label15:                                                ; preds = %label10
+  ret i32 0
+label16:                                                ; preds = %label10
+  %op17 = load i32, i32* %op2
+  %op18 = mul i32 %op17, 11
+  %op19 = load i32, i32* %op3
+  %op20 = add i32 %op18, %op19
+  %op21 = icmp slt i32 %op20, 0
+  br i1 %op21, label %label22, label %label23
+label22:                                                ; preds = %label16
+  call void @neg_idx_except()
+  ret i32 0
+label23:                                                ; preds = %label16
+  %op24 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op20
+  %op25 = load i32, i32* %op24
+  %op26 = icmp sge i32 %op25, 0
+  %op27 = zext i1 %op26 to i32
+  %op28 = icmp ne i32 %op27, 0
+  br i1 %op28, label %label29, label %label35
+label29:                                                ; preds = %label23
+  %op30 = load i32, i32* %op2
+  %op31 = mul i32 %op30, 11
+  %op32 = load i32, i32* %op3
+  %op33 = add i32 %op31, %op32
+  %op34 = icmp slt i32 %op33, 0
+  br i1 %op34, label %label40, label %label41
+label35:                                                ; preds = %label23
+  %op36 = load i32, i32* %op3
+  %op37 = load i32, i32* %op2
+  %op38 = sub i32 %op37, 1
+  %op39 = icmp slt i32 %op38, 0
+  br i1 %op39, label %label44, label %label45
+label40:                                                ; preds = %label29
+  call void @neg_idx_except()
+  ret i32 0
+label41:                                                ; preds = %label29
+  %op42 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op33
+  %op43 = load i32, i32* %op42
+  ret i32 %op43
+label44:                                                ; preds = %label35
+  call void @neg_idx_except()
+  ret i32 0
+label45:                                                ; preds = %label35
+  %op46 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 %op38
+  %op47 = load i32, i32* %op46
+  %op48 = icmp slt i32 %op36, %op47
+  %op49 = zext i1 %op48 to i32
+  %op50 = icmp ne i32 %op49, 0
+  br i1 %op50, label %label51, label %label63
+label51:                                                ; preds = %label45
+  %op52 = load i32, i32* %op2
+  %op53 = sub i32 %op52, 1
+  %op54 = load i32, i32* %op3
+  %op55 = call i32 @knapsack(i32 %op53, i32 %op54)
+  store i32 %op55, i32* %op4
+  br label %label56
+label56:                                                ; preds = %label51, %label84
+  %op57 = load i32, i32* %op4
+  %op58 = load i32, i32* %op2
+  %op59 = mul i32 %op58, 11
+  %op60 = load i32, i32* %op3
+  %op61 = add i32 %op59, %op60
+  %op62 = icmp slt i32 %op61, 0
+  br i1 %op62, label %label89, label %label90
+label63:                                                ; preds = %label45
+  %op64 = load i32, i32* %op2
+  %op65 = sub i32 %op64, 1
+  %op66 = load i32, i32* %op3
+  %op67 = call i32 @knapsack(i32 %op65, i32 %op66)
+  %op68 = load i32, i32* %op2
+  %op69 = sub i32 %op68, 1
+  %op70 = load i32, i32* %op3
+  %op71 = load i32, i32* %op2
+  %op72 = sub i32 %op71, 1
+  %op73 = icmp slt i32 %op72, 0
+  br i1 %op73, label %label74, label %label75
+label74:                                                ; preds = %label63
+  call void @neg_idx_except()
+  ret i32 0
+label75:                                                ; preds = %label63
+  %op76 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 %op72
+  %op77 = load i32, i32* %op76
+  %op78 = sub i32 %op70, %op77
+  %op79 = call i32 @knapsack(i32 %op69, i32 %op78)
+  %op80 = load i32, i32* %op2
+  %op81 = sub i32 %op80, 1
+  %op82 = icmp slt i32 %op81, 0
+  br i1 %op82, label %label83, label %label84
+label83:                                                ; preds = %label75
+  call void @neg_idx_except()
+  ret i32 0
+label84:                                                ; preds = %label75
+  %op85 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 %op81
+  %op86 = load i32, i32* %op85
+  %op87 = add i32 %op79, %op86
+  %op88 = call i32 @max(i32 %op67, i32 %op87)
+  store i32 %op88, i32* %op4
+  br label %label56
+label89:                                                ; preds = %label56
+  call void @neg_idx_except()
+  ret i32 0
+label90:                                                ; preds = %label56
+  %op91 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op61
+  store i32 %op57, i32* %op91
+  %op92 = load i32, i32* %op4
+  ret i32 %op92
+}
+define i32 @main() {
+label_entry:
+  %op0 = alloca i32
+  store i32 0, i32* %op0
+  store i32 5, i32* @n
+  store i32 10, i32* @m
+  %op1 = icmp slt i32 0, 0
+  br i1 %op1, label %label2, label %label3
+label2:                                                ; preds = %label_entry
+  call void @neg_idx_except()
+  ret i32 0
+label3:                                                ; preds = %label_entry
+  %op4 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 0
+  store i32 2, i32* %op4
+  %op5 = icmp slt i32 1, 0
+  br i1 %op5, label %label6, label %label7
+label6:                                                ; preds = %label3
+  call void @neg_idx_except()
+  ret i32 0
+label7:                                                ; preds = %label3
+  %op8 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 1
+  store i32 2, i32* %op8
+  %op9 = icmp slt i32 2, 0
+  br i1 %op9, label %label10, label %label11
+label10:                                                ; preds = %label7
+  call void @neg_idx_except()
+  ret i32 0
+label11:                                                ; preds = %label7
+  %op12 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 2
+  store i32 6, i32* %op12
+  %op13 = icmp slt i32 3, 0
+  br i1 %op13, label %label14, label %label15
+label14:                                                ; preds = %label11
+  call void @neg_idx_except()
+  ret i32 0
+label15:                                                ; preds = %label11
+  %op16 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 3
+  store i32 5, i32* %op16
+  %op17 = icmp slt i32 4, 0
+  br i1 %op17, label %label18, label %label19
+label18:                                                ; preds = %label15
+  call void @neg_idx_except()
+  ret i32 0
+label19:                                                ; preds = %label15
+  %op20 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 4
+  store i32 4, i32* %op20
+  %op21 = icmp slt i32 0, 0
+  br i1 %op21, label %label22, label %label23
+label22:                                                ; preds = %label19
+  call void @neg_idx_except()
+  ret i32 0
+label23:                                                ; preds = %label19
+  %op24 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 0
+  store i32 6, i32* %op24
+  %op25 = icmp slt i32 1, 0
+  br i1 %op25, label %label26, label %label27
+label26:                                                ; preds = %label23
+  call void @neg_idx_except()
+  ret i32 0
+label27:                                                ; preds = %label23
+  %op28 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 1
+  store i32 3, i32* %op28
+  %op29 = icmp slt i32 2, 0
+  br i1 %op29, label %label30, label %label31
+label30:                                                ; preds = %label27
+  call void @neg_idx_except()
+  ret i32 0
+label31:                                                ; preds = %label27
+  %op32 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 2
+  store i32 5, i32* %op32
+  %op33 = icmp slt i32 3, 0
+  br i1 %op33, label %label34, label %label35
+label34:                                                ; preds = %label31
+  call void @neg_idx_except()
+  ret i32 0
+label35:                                                ; preds = %label31
+  %op36 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 3
+  store i32 4, i32* %op36
+  %op37 = icmp slt i32 4, 0
+  br i1 %op37, label %label38, label %label39
+label38:                                                ; preds = %label35
+  call void @neg_idx_except()
+  ret i32 0
+label39:                                                ; preds = %label35
+  %op40 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 4
+  store i32 6, i32* %op40
+  br label %label41
+label41:                                                ; preds = %label39, %label55
+  %op42 = load i32, i32* %op0
+  %op43 = icmp slt i32 %op42, 66
+  %op44 = zext i1 %op43 to i32
+  %op45 = icmp ne i32 %op44, 0
+  br i1 %op45, label %label46, label %label50
+label46:                                                ; preds = %label41
+  %op47 = sub i32 0, 1
+  %op48 = load i32, i32* %op0
+  %op49 = icmp slt i32 %op48, 0
+  br i1 %op49, label %label54, label %label55
+label50:                                                ; preds = %label41
+  %op51 = load i32, i32* @n
+  %op52 = load i32, i32* @m
+  %op53 = call i32 @knapsack(i32 %op51, i32 %op52)
+  call void @output(i32 %op53)
+  ret i32 0
+label54:                                                ; preds = %label46
+  call void @neg_idx_except()
+  ret i32 0
+label55:                                                ; preds = %label46
+  %op56 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op48
+  store i32 %op47, i32* %op56
+  %op57 = load i32, i32* %op0
+  %op58 = add i32 %op57, 1
+  store i32 %op58, i32* %op0
+  br label %label41
+}

tests/3-codegen/autogen/output/13-complex.out

+15
+0

tests/3-codegen/autogen/output/13-complex.s

+# Global variables
+	.text
+	.section .bss, "aw", @nobits
+	.globl n
+	.type n, @object
+	.size n, 4
+n:
+	.space 4
+	.globl m
+	.type m, @object
+	.size m, 4
+m:
+	.space 4
+	.globl w
+	.type w, @object
+	.size w, 20
+w:
+	.space 20
+	.globl v
+	.type v, @object
+	.size v, 20
+v:
+	.space 20
+	.globl dp
+	.type dp, @object
+	.size dp, 264
+dp:
+	.space 264
+	.text
+	.globl max
+	.type max, @function
+max:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+	st.w $a0, $fp, -20
+	st.w $a1, $fp, -24
+.max_label_entry:
+# %op2 = alloca i32
+	addi.d $t0, $fp, -36
+	st.d $t0, $fp, -32
+# store i32 %arg0, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# %op3 = alloca i32
+	addi.d $t0, $fp, -48
+	st.d $t0, $fp, -44
+# store i32 %arg1, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t1, $fp, -24
+	st.w $t1, $t0, 0
+# %op4 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -52
+# %op5 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op6 = icmp sgt i32 %op4, %op5
+	ld.w $t0, $fp, -52
+	ld.w $t1, $fp, -56
+	slt $t2, $t1, $t0
+	st.b $t2, $fp, -57
+# %op7 = zext i1 %op6 to i32
+	ld.b $t0, $fp, -57
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -61
+# %op8 = icmp ne i32 %op7, 0
+	ld.w $t0, $fp, -61
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -62
+# br i1 %op8, label %label9, label %label12
+	ld.b $t0, $fp, -62
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .max_label9
+	b .max_label12
+.max_label9:
+# %op10 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -66
+# ret i32 %op10
+	ld.w $a0, $fp, -66
+	b max_exit
+.max_label11:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b max_exit
+.max_label12:
+# %op13 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -70
+# ret i32 %op13
+	ld.w $a0, $fp, -70
+	b max_exit
+max_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl knapsack
+	.type knapsack, @function
+knapsack:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -320
+	st.w $a0, $fp, -20
+	st.w $a1, $fp, -24
+.knapsack_label_entry:
+# %op2 = alloca i32
+	addi.d $t0, $fp, -36
+	st.d $t0, $fp, -32
+# store i32 %arg0, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# %op3 = alloca i32
+	addi.d $t0, $fp, -48
+	st.d $t0, $fp, -44
+# store i32 %arg1, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t1, $fp, -24
+	st.w $t1, $t0, 0
+# %op4 = alloca i32
+	addi.d $t0, $fp, -60
+	st.d $t0, $fp, -56
+# %op5 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -64
+# %op6 = icmp sle i32 %op5, 0
+	ld.w $t0, $fp, -64
+	addi.w $t1, $zero, 0
+	slt $t2, $t1, $t0
+	xori $t2, $t2, 1
+	st.b $t2, $fp, -65
+# %op7 = zext i1 %op6 to i32
+	ld.b $t0, $fp, -65
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -69
+# %op8 = icmp ne i32 %op7, 0
+	ld.w $t0, $fp, -69
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -70
+# br i1 %op8, label %label9, label %label10
+	ld.b $t0, $fp, -70
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label9
+	b .knapsack_label10
+.knapsack_label9:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label10:
+# %op11 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -74
+# %op12 = icmp eq i32 %op11, 0
+	ld.w $t0, $fp, -74
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltui $t2, $t2, 1
+	st.b $t2, $fp, -75
+# %op13 = zext i1 %op12 to i32
+	ld.b $t0, $fp, -75
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -79
+# %op14 = icmp ne i32 %op13, 0
+	ld.w $t0, $fp, -79
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -80
+# br i1 %op14, label %label15, label %label16
+	ld.b $t0, $fp, -80
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label15
+	b .knapsack_label16
+.knapsack_label15:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label16:
+# %op17 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -84
+# %op18 = mul i32 %op17, 11
+	ld.w $t0, $fp, -84
+	addi.w $t1, $zero, 11
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -88
+# %op19 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -92
+# %op20 = add i32 %op18, %op19
+	ld.w $t0, $fp, -88
+	ld.w $t1, $fp, -92
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -96
+# %op21 = icmp slt i32 %op20, 0
+	ld.w $t0, $fp, -96
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -97
+# br i1 %op21, label %label22, label %label23
+	ld.b $t0, $fp, -97
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label22
+	b .knapsack_label23
+.knapsack_label22:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label23:
+# %op24 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op20
+	la.local $t0, dp
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 264
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -96
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -105
+# %op25 = load i32, i32* %op24
+	ld.d $t0, $fp, -105
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -109
+# %op26 = icmp sge i32 %op25, 0
+	ld.w $t0, $fp, -109
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	xori $t2, $t2, 1
+	st.b $t2, $fp, -110
+# %op27 = zext i1 %op26 to i32
+	ld.b $t0, $fp, -110
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -114
+# %op28 = icmp ne i32 %op27, 0
+	ld.w $t0, $fp, -114
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -115
+# br i1 %op28, label %label29, label %label35
+	ld.b $t0, $fp, -115
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label29
+	b .knapsack_label35
+.knapsack_label29:
+# %op30 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -119
+# %op31 = mul i32 %op30, 11
+	ld.w $t0, $fp, -119
+	addi.w $t1, $zero, 11
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -123
+# %op32 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -127
+# %op33 = add i32 %op31, %op32
+	ld.w $t0, $fp, -123
+	ld.w $t1, $fp, -127
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -131
+# %op34 = icmp slt i32 %op33, 0
+	ld.w $t0, $fp, -131
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -132
+# br i1 %op34, label %label40, label %label41
+	ld.b $t0, $fp, -132
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label40
+	b .knapsack_label41
+.knapsack_label35:
+# %op36 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -136
+# %op37 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -140
+# %op38 = sub i32 %op37, 1
+	ld.w $t0, $fp, -140
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -144
+# %op39 = icmp slt i32 %op38, 0
+	ld.w $t0, $fp, -144
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -145
+# br i1 %op39, label %label44, label %label45
+	ld.b $t0, $fp, -145
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label44
+	b .knapsack_label45
+.knapsack_label40:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label41:
+# %op42 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op33
+	la.local $t0, dp
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 264
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -131
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -153
+# %op43 = load i32, i32* %op42
+	ld.d $t0, $fp, -153
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -157
+# ret i32 %op43
+	ld.w $a0, $fp, -157
+	b knapsack_exit
+.knapsack_label44:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label45:
+# %op46 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 %op38
+	la.local $t0, w
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -144
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -165
+# %op47 = load i32, i32* %op46
+	ld.d $t0, $fp, -165
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -169
+# %op48 = icmp slt i32 %op36, %op47
+	ld.w $t0, $fp, -136
+	ld.w $t1, $fp, -169
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -170
+# %op49 = zext i1 %op48 to i32
+	ld.b $t0, $fp, -170
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -174
+# %op50 = icmp ne i32 %op49, 0
+	ld.w $t0, $fp, -174
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -175
+# br i1 %op50, label %label51, label %label63
+	ld.b $t0, $fp, -175
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label51
+	b .knapsack_label63
+.knapsack_label51:
+# %op52 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -179
+# %op53 = sub i32 %op52, 1
+	ld.w $t0, $fp, -179
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -183
+# %op54 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -187
+# %op55 = call i32 @knapsack(i32 %op53, i32 %op54)
+	ld.w $a0, $fp, -183
+	ld.w $a1, $fp, -187
+	bl knapsack
+	st.w $a0, $fp, -191
+# store i32 %op55, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t1, $fp, -191
+	st.w $t1, $t0, 0
+# br label %label56
+	b .knapsack_label56
+.knapsack_label56:
+# %op57 = load i32, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -195
+# %op58 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -199
+# %op59 = mul i32 %op58, 11
+	ld.w $t0, $fp, -199
+	addi.w $t1, $zero, 11
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -203
+# %op60 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -207
+# %op61 = add i32 %op59, %op60
+	ld.w $t0, $fp, -203
+	ld.w $t1, $fp, -207
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -211
+# %op62 = icmp slt i32 %op61, 0
+	ld.w $t0, $fp, -211
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -212
+# br i1 %op62, label %label89, label %label90
+	ld.b $t0, $fp, -212
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label89
+	b .knapsack_label90
+.knapsack_label63:
+# %op64 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -216
+# %op65 = sub i32 %op64, 1
+	ld.w $t0, $fp, -216
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -220
+# %op66 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -224
+# %op67 = call i32 @knapsack(i32 %op65, i32 %op66)
+	ld.w $a0, $fp, -220
+	ld.w $a1, $fp, -224
+	bl knapsack
+	st.w $a0, $fp, -228
+# %op68 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -232
+# %op69 = sub i32 %op68, 1
+	ld.w $t0, $fp, -232
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -236
+# %op70 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -240
+# %op71 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -244
+# %op72 = sub i32 %op71, 1
+	ld.w $t0, $fp, -244
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -248
+# %op73 = icmp slt i32 %op72, 0
+	ld.w $t0, $fp, -248
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -249
+# br i1 %op73, label %label74, label %label75
+	ld.b $t0, $fp, -249
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label74
+	b .knapsack_label75
+.knapsack_label74:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label75:
+# %op76 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 %op72
+	la.local $t0, w
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -248
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -257
+# %op77 = load i32, i32* %op76
+	ld.d $t0, $fp, -257
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -261
+# %op78 = sub i32 %op70, %op77
+	ld.w $t0, $fp, -240
+	ld.w $t1, $fp, -261
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -265
+# %op79 = call i32 @knapsack(i32 %op69, i32 %op78)
+	ld.w $a0, $fp, -236
+	ld.w $a1, $fp, -265
+	bl knapsack
+	st.w $a0, $fp, -269
+# %op80 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -273
+# %op81 = sub i32 %op80, 1
+	ld.w $t0, $fp, -273
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -277
+# %op82 = icmp slt i32 %op81, 0
+	ld.w $t0, $fp, -277
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -278
+# br i1 %op82, label %label83, label %label84
+	ld.b $t0, $fp, -278
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .knapsack_label83
+	b .knapsack_label84
+.knapsack_label83:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label84:
+# %op85 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 %op81
+	la.local $t0, v
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -277
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -286
+# %op86 = load i32, i32* %op85
+	ld.d $t0, $fp, -286
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -290
+# %op87 = add i32 %op79, %op86
+	ld.w $t0, $fp, -269
+	ld.w $t1, $fp, -290
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -294
+# %op88 = call i32 @max(i32 %op67, i32 %op87)
+	ld.w $a0, $fp, -228
+	ld.w $a1, $fp, -294
+	bl max
+	st.w $a0, $fp, -298
+# store i32 %op88, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t1, $fp, -298
+	st.w $t1, $t0, 0
+# br label %label56
+	b .knapsack_label56
+.knapsack_label89:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b knapsack_exit
+.knapsack_label90:
+# %op91 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op61
+	la.local $t0, dp
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 264
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -211
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -306
+# store i32 %op57, i32* %op91
+	ld.d $t0, $fp, -306
+	ld.w $t1, $fp, -195
+	st.w $t1, $t0, 0
+# %op92 = load i32, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -310
+# ret i32 %op92
+	ld.w $a0, $fp, -310
+	b knapsack_exit
+knapsack_exit:
+	addi.d $sp, $sp, 320
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -176
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# store i32 0, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# store i32 5, i32* @n
+	la.local $t0, n
+	addi.w $t1, $zero, 5
+	st.w $t1, $t0, 0
+# store i32 10, i32* @m
+	la.local $t0, m
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# %op1 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -29
+# br i1 %op1, label %label2, label %label3
+	ld.b $t0, $fp, -29
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label2
+	b .main_label3
+.main_label2:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label3:
+# %op4 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 0
+	la.local $t0, w
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -37
+# store i32 2, i32* %op4
+	ld.d $t0, $fp, -37
+	addi.w $t1, $zero, 2
+	st.w $t1, $t0, 0
+# %op5 = icmp slt i32 1, 0
+	addi.w $t0, $zero, 1
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -38
+# br i1 %op5, label %label6, label %label7
+	ld.b $t0, $fp, -38
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label6
+	b .main_label7
+.main_label6:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label7:
+# %op8 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 1
+	la.local $t0, w
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 1
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -46
+# store i32 2, i32* %op8
+	ld.d $t0, $fp, -46
+	addi.w $t1, $zero, 2
+	st.w $t1, $t0, 0
+# %op9 = icmp slt i32 2, 0
+	addi.w $t0, $zero, 2
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -47
+# br i1 %op9, label %label10, label %label11
+	ld.b $t0, $fp, -47
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label10
+	b .main_label11
+.main_label10:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label11:
+# %op12 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 2
+	la.local $t0, w
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 2
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -55
+# store i32 6, i32* %op12
+	ld.d $t0, $fp, -55
+	addi.w $t1, $zero, 6
+	st.w $t1, $t0, 0
+# %op13 = icmp slt i32 3, 0
+	addi.w $t0, $zero, 3
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -56
+# br i1 %op13, label %label14, label %label15
+	ld.b $t0, $fp, -56
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label14
+	b .main_label15
+.main_label14:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label15:
+# %op16 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 3
+	la.local $t0, w
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 3
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -64
+# store i32 5, i32* %op16
+	ld.d $t0, $fp, -64
+	addi.w $t1, $zero, 5
+	st.w $t1, $t0, 0
+# %op17 = icmp slt i32 4, 0
+	addi.w $t0, $zero, 4
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -65
+# br i1 %op17, label %label18, label %label19
+	ld.b $t0, $fp, -65
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label18
+	b .main_label19
+.main_label18:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label19:
+# %op20 = getelementptr [5 x i32], [5 x i32]* @w, i32 0, i32 4
+	la.local $t0, w
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 4
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -73
+# store i32 4, i32* %op20
+	ld.d $t0, $fp, -73
+	addi.w $t1, $zero, 4
+	st.w $t1, $t0, 0
+# %op21 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -74
+# br i1 %op21, label %label22, label %label23
+	ld.b $t0, $fp, -74
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label22
+	b .main_label23
+.main_label22:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label23:
+# %op24 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 0
+	la.local $t0, v
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -82
+# store i32 6, i32* %op24
+	ld.d $t0, $fp, -82
+	addi.w $t1, $zero, 6
+	st.w $t1, $t0, 0
+# %op25 = icmp slt i32 1, 0
+	addi.w $t0, $zero, 1
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -83
+# br i1 %op25, label %label26, label %label27
+	ld.b $t0, $fp, -83
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label26
+	b .main_label27
+.main_label26:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label27:
+# %op28 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 1
+	la.local $t0, v
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 1
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -91
+# store i32 3, i32* %op28
+	ld.d $t0, $fp, -91
+	addi.w $t1, $zero, 3
+	st.w $t1, $t0, 0
+# %op29 = icmp slt i32 2, 0
+	addi.w $t0, $zero, 2
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -92
+# br i1 %op29, label %label30, label %label31
+	ld.b $t0, $fp, -92
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label30
+	b .main_label31
+.main_label30:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label31:
+# %op32 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 2
+	la.local $t0, v
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 2
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -100
+# store i32 5, i32* %op32
+	ld.d $t0, $fp, -100
+	addi.w $t1, $zero, 5
+	st.w $t1, $t0, 0
+# %op33 = icmp slt i32 3, 0
+	addi.w $t0, $zero, 3
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -101
+# br i1 %op33, label %label34, label %label35
+	ld.b $t0, $fp, -101
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label34
+	b .main_label35
+.main_label34:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label35:
+# %op36 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 3
+	la.local $t0, v
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 3
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -109
+# store i32 4, i32* %op36
+	ld.d $t0, $fp, -109
+	addi.w $t1, $zero, 4
+	st.w $t1, $t0, 0
+# %op37 = icmp slt i32 4, 0
+	addi.w $t0, $zero, 4
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -110
+# br i1 %op37, label %label38, label %label39
+	ld.b $t0, $fp, -110
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label38
+	b .main_label39
+.main_label38:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label39:
+# %op40 = getelementptr [5 x i32], [5 x i32]* @v, i32 0, i32 4
+	la.local $t0, v
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 20
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 4
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -118
+# store i32 6, i32* %op40
+	ld.d $t0, $fp, -118
+	addi.w $t1, $zero, 6
+	st.w $t1, $t0, 0
+# br label %label41
+	b .main_label41
+.main_label41:
+# %op42 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -122
+# %op43 = icmp slt i32 %op42, 66
+	ld.w $t0, $fp, -122
+	addi.w $t1, $zero, 66
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -123
+# %op44 = zext i1 %op43 to i32
+	ld.b $t0, $fp, -123
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -127
+# %op45 = icmp ne i32 %op44, 0
+	ld.w $t0, $fp, -127
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -128
+# br i1 %op45, label %label46, label %label50
+	ld.b $t0, $fp, -128
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label46
+	b .main_label50
+.main_label46:
+# %op47 = sub i32 0, 1
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -132
+# %op48 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -136
+# %op49 = icmp slt i32 %op48, 0
+	ld.w $t0, $fp, -136
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -137
+# br i1 %op49, label %label54, label %label55
+	ld.b $t0, $fp, -137
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label54
+	b .main_label55
+.main_label50:
+# %op51 = load i32, i32* @n
+	la.local $t0, n
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -141
+# %op52 = load i32, i32* @m
+	la.local $t0, m
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -145
+# %op53 = call i32 @knapsack(i32 %op51, i32 %op52)
+	ld.w $a0, $fp, -141
+	ld.w $a1, $fp, -145
+	bl knapsack
+	st.w $a0, $fp, -149
+# call void @output(i32 %op53)
+	ld.w $a0, $fp, -149
+	bl output
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label54:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label55:
+# %op56 = getelementptr [66 x i32], [66 x i32]* @dp, i32 0, i32 %op48
+	la.local $t0, dp
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 264
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -136
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -157
+# store i32 %op47, i32* %op56
+	ld.d $t0, $fp, -157
+	ld.w $t1, $fp, -132
+	st.w $t1, $t0, 0
+# %op57 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -161
+# %op58 = add i32 %op57, 1
+	ld.w $t0, $fp, -161
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -165
+# store i32 %op58, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -165
+	st.w $t1, $t0, 0
+# br label %label41
+	b .main_label41
+main_exit:
+	addi.d $sp, $sp, 176
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/13-if_stmt.out

+0

tests/3-codegen/autogen/output/13-if_stmt.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -48
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# store i32 2, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 2
+	st.w $t1, $t0, 0
+# %op1 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -32
+# %op2 = icmp ne i32 %op1, 0
+	ld.w $t0, $fp, -32
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -33
+# br i1 %op2, label %label3, label %label4
+	ld.b $t0, $fp, -33
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label3
+	b .main_label4
+.main_label3:
+# store i32 3, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 3
+	st.w $t1, $t0, 0
+# br label %label4
+	b .main_label4
+.main_label4:
+# store i32 4, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 4
+	st.w $t1, $t0, 0
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 48
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/14-while_stmt.out

+0

tests/3-codegen/autogen/output/14-while_stmt.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -48
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# store i32 10, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# br label %label1
+	b .main_label1
+.main_label1:
+# %op2 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -32
+# %op3 = icmp ne i32 %op2, 0
+	ld.w $t0, $fp, -32
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -33
+# br i1 %op3, label %label4, label %label7
+	ld.b $t0, $fp, -33
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label4
+	b .main_label7
+.main_label4:
+# %op5 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -37
+# %op6 = sub i32 %op5, 1
+	ld.w $t0, $fp, -37
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -41
+# store i32 %op6, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -41
+	st.w $t1, $t0, 0
+# br label %label1
+	b .main_label1
+.main_label7:
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 48
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/15-if_while.out

+45

tests/3-codegen/autogen/output/15-if_while.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -112
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# %op2 = alloca i32
+	addi.d $t0, $fp, -52
+	st.d $t0, $fp, -48
+# store i32 0, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# store i32 0, i32* %op2
+	ld.d $t0, $fp, -48
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# store i32 10, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# br label %label3
+	b .main_label3
+.main_label3:
+# %op4 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op5 = icmp ne i32 %op4, 0
+	ld.w $t0, $fp, -56
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -57
+# br i1 %op5, label %label6, label %label13
+	ld.b $t0, $fp, -57
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label6
+	b .main_label13
+.main_label6:
+# %op7 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -61
+# %op8 = sub i32 %op7, 1
+	ld.w $t0, $fp, -61
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -65
+# store i32 %op8, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -65
+	st.w $t1, $t0, 0
+# %op9 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -69
+# %op10 = icmp slt i32 %op9, 5
+	ld.w $t0, $fp, -69
+	addi.w $t1, $zero, 5
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -70
+# %op11 = zext i1 %op10 to i32
+	ld.b $t0, $fp, -70
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -74
+# %op12 = icmp ne i32 %op11, 0
+	ld.w $t0, $fp, -74
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -75
+# br i1 %op12, label %label17, label %label22
+	ld.b $t0, $fp, -75
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label17
+	b .main_label22
+.main_label13:
+# %op14 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -79
+# %op15 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -83
+# %op16 = add i32 %op14, %op15
+	ld.w $t0, $fp, -79
+	ld.w $t1, $fp, -83
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -87
+# ret i32 %op16
+	ld.w $a0, $fp, -87
+	b main_exit
+.main_label17:
+# %op18 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -91
+# %op19 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -95
+# %op20 = add i32 %op18, %op19
+	ld.w $t0, $fp, -91
+	ld.w $t1, $fp, -95
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -99
+# store i32 %op20, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t1, $fp, -99
+	st.w $t1, $t0, 0
+# br label %label21
+	b .main_label21
+.main_label21:
+# br label %label3
+	b .main_label3
+.main_label22:
+# %op23 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -103
+# %op24 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -107
+# %op25 = add i32 %op23, %op24
+	ld.w $t0, $fp, -103
+	ld.w $t1, $fp, -107
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -111
+# store i32 %op25, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t1, $fp, -111
+	st.w $t1, $t0, 0
+# br label %label21
+	b .main_label21
+main_exit:
+	addi.d $sp, $sp, 112
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/16-if_chain.out

+3

tests/3-codegen/autogen/output/16-if_chain.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# %op2 = alloca i32
+	addi.d $t0, $fp, -52
+	st.d $t0, $fp, -48
+# store i32 0, i32* %op2
+	ld.d $t0, $fp, -48
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# store i32 2, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 2
+	st.w $t1, $t0, 0
+# store i32 1, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 1
+	st.w $t1, $t0, 0
+# %op3 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op4 = icmp ne i32 %op3, 0
+	ld.w $t0, $fp, -56
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -57
+# br i1 %op4, label %label5, label %label8
+	ld.b $t0, $fp, -57
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label5
+	b .main_label8
+.main_label5:
+# %op6 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -61
+# %op7 = icmp ne i32 %op6, 0
+	ld.w $t0, $fp, -61
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -62
+# br i1 %op7, label %label10, label %label12
+	ld.b $t0, $fp, -62
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label10
+	b .main_label12
+.main_label8:
+# %op9 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -66
+# ret i32 %op9
+	ld.w $a0, $fp, -66
+	b main_exit
+.main_label10:
+# store i32 4, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 4
+	st.w $t1, $t0, 0
+# br label %label11
+	b .main_label11
+.main_label11:
+# br label %label8
+	b .main_label8
+.main_label12:
+# store i32 3, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 3
+	st.w $t1, $t0, 0
+# br label %label11
+	b .main_label11
+main_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/17-while_chain.out

+0

tests/3-codegen/autogen/output/17-while_chain.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -96
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# store i32 10, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# br label %label2
+	b .main_label2
+.main_label2:
+# %op3 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -44
+# %op4 = icmp ne i32 %op3, 0
+	ld.w $t0, $fp, -44
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -45
+# br i1 %op4, label %label5, label %label9
+	ld.b $t0, $fp, -45
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label5
+	b .main_label9
+.main_label5:
+# %op6 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -49
+# %op7 = sub i32 %op6, 1
+	ld.w $t0, $fp, -49
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -53
+# store i32 %op7, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -53
+	st.w $t1, $t0, 0
+# %op8 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -57
+# store i32 %op8, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t1, $fp, -57
+	st.w $t1, $t0, 0
+# br label %label13
+	b .main_label13
+.main_label9:
+# %op10 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -61
+# %op11 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -65
+# %op12 = add i32 %op10, %op11
+	ld.w $t0, $fp, -61
+	ld.w $t1, $fp, -65
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -69
+# ret i32 %op12
+	ld.w $a0, $fp, -69
+	b main_exit
+.main_label13:
+# %op14 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -73
+# %op15 = icmp ne i32 %op14, 0
+	ld.w $t0, $fp, -73
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -74
+# br i1 %op15, label %label16, label %label19
+	ld.b $t0, $fp, -74
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label16
+	b .main_label19
+.main_label16:
+# %op17 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -78
+# %op18 = sub i32 %op17, 1
+	ld.w $t0, $fp, -78
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -82
+# store i32 %op18, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t1, $fp, -82
+	st.w $t1, $t0, 0
+# br label %label13
+	b .main_label13
+.main_label19:
+# br label %label2
+	b .main_label2
+main_exit:
+	addi.d $sp, $sp, 96
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/18-global_var.out

+10

tests/3-codegen/autogen/output/18-global_var.s

+# Global variables
+	.text
+	.section .bss, "aw", @nobits
+	.globl a
+	.type a, @object
+	.size a, 4
+a:
+	.space 4
+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.main_label_entry:
+# store i32 10, i32* @a
+	la.local $t0, a
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# %op0 = load i32, i32* @a
+	la.local $t0, a
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -20
+# ret i32 %op0
+	ld.w $a0, $fp, -20
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/19-global_local_var.out

+20

tests/3-codegen/autogen/output/19-global_local_var.s

+# Global variables
+	.text
+	.section .bss, "aw", @nobits
+	.globl a
+	.type a, @object
+	.size a, 4
+a:
+	.space 4
+	.text
+	.globl GlobalAssign
+	.type GlobalAssign, @function
+GlobalAssign:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -16
+.GlobalAssign_label_entry:
+# store i32 10, i32* @a
+	la.local $t0, a
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# ret void
+	addi.w $a0, $zero, 0
+	b GlobalAssign_exit
+GlobalAssign_exit:
+	addi.d $sp, $sp, 16
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# call void @GlobalAssign()
+	bl GlobalAssign
+# store i32 20, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 20
+	st.w $t1, $t0, 0
+# %op1 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -32
+# ret i32 %op1
+	ld.w $a0, $fp, -32
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/2-calculate.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/2-calculate.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  %op0 = alloca i32
+  %op1 = alloca i32
+  %op2 = alloca i32
+  store i32 23, i32* %op0
+  store i32 25, i32* %op1
+  store i32 4, i32* %op2
+  %op3 = load i32, i32* %op0
+  %op4 = load i32, i32* %op1
+  %op5 = load i32, i32* %op2
+  %op6 = mul i32 %op4, %op5
+  %op7 = add i32 %op3, %op6
+  ret i32 %op7
+}

tests/3-codegen/autogen/output/2-calculate.out

+123

tests/3-codegen/autogen/output/2-calculate.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# %op2 = alloca i32
+	addi.d $t0, $fp, -52
+	st.d $t0, $fp, -48
+# store i32 23, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 23
+	st.w $t1, $t0, 0
+# store i32 25, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 25
+	st.w $t1, $t0, 0
+# store i32 4, i32* %op2
+	ld.d $t0, $fp, -48
+	addi.w $t1, $zero, 4
+	st.w $t1, $t0, 0
+# %op3 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op4 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -60
+# %op5 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -64
+# %op6 = mul i32 %op4, %op5
+	ld.w $t0, $fp, -60
+	ld.w $t1, $fp, -64
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -68
+# %op7 = add i32 %op3, %op6
+	ld.w $t0, $fp, -56
+	ld.w $t1, $fp, -68
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -72
+# ret i32 %op7
+	ld.w $a0, $fp, -72
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/2-decl_int.out

+0

tests/3-codegen/autogen/output/2-decl_int.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/20-gcd_array.out

+18

tests/3-codegen/autogen/output/20-gcd_array.s

+# Global variables
+	.text
+	.section .bss, "aw", @nobits
+	.globl x
+	.type x, @object
+	.size x, 4
+x:
+	.space 4
+	.globl y
+	.type y, @object
+	.size y, 4
+y:
+	.space 4
+	.text
+	.globl gcd
+	.type gcd, @function
+gcd:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -112
+	st.w $a0, $fp, -20
+	st.w $a1, $fp, -24
+.gcd_label_entry:
+# %op2 = alloca i32
+	addi.d $t0, $fp, -36
+	st.d $t0, $fp, -32
+# store i32 %arg0, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# %op3 = alloca i32
+	addi.d $t0, $fp, -48
+	st.d $t0, $fp, -44
+# store i32 %arg1, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t1, $fp, -24
+	st.w $t1, $t0, 0
+# %op4 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -52
+# %op5 = icmp eq i32 %op4, 0
+	ld.w $t0, $fp, -52
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltui $t2, $t2, 1
+	st.b $t2, $fp, -53
+# %op6 = zext i1 %op5 to i32
+	ld.b $t0, $fp, -53
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -57
+# %op7 = icmp ne i32 %op6, 0
+	ld.w $t0, $fp, -57
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -58
+# br i1 %op7, label %label8, label %label11
+	ld.b $t0, $fp, -58
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .gcd_label8
+	b .gcd_label11
+.gcd_label8:
+# %op9 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -62
+# ret i32 %op9
+	ld.w $a0, $fp, -62
+	b gcd_exit
+.gcd_label10:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b gcd_exit
+.gcd_label11:
+# %op12 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -66
+# %op13 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -70
+# %op14 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -74
+# %op15 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -78
+# %op16 = sdiv i32 %op14, %op15
+	ld.w $t0, $fp, -74
+	ld.w $t1, $fp, -78
+	div.w $t2, $t0, $t1
+	st.w $t2, $fp, -82
+# %op17 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -86
+# %op18 = mul i32 %op16, %op17
+	ld.w $t0, $fp, -82
+	ld.w $t1, $fp, -86
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -90
+# %op19 = sub i32 %op13, %op18
+	ld.w $t0, $fp, -70
+	ld.w $t1, $fp, -90
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -94
+# %op20 = call i32 @gcd(i32 %op12, i32 %op19)
+	ld.w $a0, $fp, -66
+	ld.w $a1, $fp, -94
+	bl gcd
+	st.w $a0, $fp, -98
+# ret i32 %op20
+	ld.w $a0, $fp, -98
+	b gcd_exit
+gcd_exit:
+	addi.d $sp, $sp, 112
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl funArray
+	.type funArray, @function
+funArray:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -192
+	st.d $a0, $fp, -24
+	st.d $a1, $fp, -32
+.funArray_label_entry:
+# %op2 = alloca i32*
+	addi.d $t0, $fp, -48
+	st.d $t0, $fp, -40
+# store i32* %arg0, i32** %op2
+	ld.d $t0, $fp, -40
+	ld.d $t1, $fp, -24
+	st.d $t1, $t0, 0
+# %op3 = alloca i32*
+	addi.d $t0, $fp, -64
+	st.d $t0, $fp, -56
+# store i32* %arg1, i32** %op3
+	ld.d $t0, $fp, -56
+	ld.d $t1, $fp, -32
+	st.d $t1, $t0, 0
+# %op4 = alloca i32
+	addi.d $t0, $fp, -76
+	st.d $t0, $fp, -72
+# %op5 = alloca i32
+	addi.d $t0, $fp, -88
+	st.d $t0, $fp, -84
+# %op6 = alloca i32
+	addi.d $t0, $fp, -100
+	st.d $t0, $fp, -96
+# %op7 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -101
+# br i1 %op7, label %label8, label %label9
+	ld.b $t0, $fp, -101
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .funArray_label8
+	b .funArray_label9
+.funArray_label8:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b funArray_exit
+.funArray_label9:
+# %op10 = load i32*, i32** %op2
+	ld.d $t0, $fp, -40
+	ld.d $t0, $t0, 0
+	st.d $t0, $fp, -109
+# %op11 = getelementptr i32, i32* %op10, i32 0
+	ld.d $t0, $fp, -109
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -117
+# %op12 = load i32, i32* %op11
+	ld.d $t0, $fp, -117
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -121
+# store i32 %op12, i32* %op4
+	ld.d $t0, $fp, -72
+	ld.w $t1, $fp, -121
+	st.w $t1, $t0, 0
+# %op13 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -122
+# br i1 %op13, label %label14, label %label15
+	ld.b $t0, $fp, -122
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .funArray_label14
+	b .funArray_label15
+.funArray_label14:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b funArray_exit
+.funArray_label15:
+# %op16 = load i32*, i32** %op3
+	ld.d $t0, $fp, -56
+	ld.d $t0, $t0, 0
+	st.d $t0, $fp, -130
+# %op17 = getelementptr i32, i32* %op16, i32 0
+	ld.d $t0, $fp, -130
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -138
+# %op18 = load i32, i32* %op17
+	ld.d $t0, $fp, -138
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -142
+# store i32 %op18, i32* %op5
+	ld.d $t0, $fp, -84
+	ld.w $t1, $fp, -142
+	st.w $t1, $t0, 0
+# %op19 = load i32, i32* %op4
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -146
+# %op20 = load i32, i32* %op5
+	ld.d $t0, $fp, -84
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -150
+# %op21 = icmp slt i32 %op19, %op20
+	ld.w $t0, $fp, -146
+	ld.w $t1, $fp, -150
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -151
+# %op22 = zext i1 %op21 to i32
+	ld.b $t0, $fp, -151
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -155
+# %op23 = icmp ne i32 %op22, 0
+	ld.w $t0, $fp, -155
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -156
+# br i1 %op23, label %label24, label %label28
+	ld.b $t0, $fp, -156
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .funArray_label24
+	b .funArray_label28
+.funArray_label24:
+# %op25 = load i32, i32* %op4
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -160
+# store i32 %op25, i32* %op6
+	ld.d $t0, $fp, -96
+	ld.w $t1, $fp, -160
+	st.w $t1, $t0, 0
+# %op26 = load i32, i32* %op5
+	ld.d $t0, $fp, -84
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -164
+# store i32 %op26, i32* %op4
+	ld.d $t0, $fp, -72
+	ld.w $t1, $fp, -164
+	st.w $t1, $t0, 0
+# %op27 = load i32, i32* %op6
+	ld.d $t0, $fp, -96
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -168
+# store i32 %op27, i32* %op5
+	ld.d $t0, $fp, -84
+	ld.w $t1, $fp, -168
+	st.w $t1, $t0, 0
+# br label %label28
+	b .funArray_label28
+.funArray_label28:
+# %op29 = load i32, i32* %op4
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -172
+# %op30 = load i32, i32* %op5
+	ld.d $t0, $fp, -84
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -176
+# %op31 = call i32 @gcd(i32 %op29, i32 %op30)
+	ld.w $a0, $fp, -172
+	ld.w $a1, $fp, -176
+	bl gcd
+	st.w $a0, $fp, -180
+# ret i32 %op31
+	ld.w $a0, $fp, -180
+	b funArray_exit
+funArray_exit:
+	addi.d $sp, $sp, 192
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -64
+.main_label_entry:
+# %op0 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -17
+# br i1 %op0, label %label1, label %label2
+	ld.b $t0, $fp, -17
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label1
+	b .main_label2
+.main_label1:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label2:
+# %op3 = getelementptr [1 x i32], [1 x i32]* @x, i32 0, i32 0
+	la.local $t0, x
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -25
+# store i32 90, i32* %op3
+	ld.d $t0, $fp, -25
+	addi.w $t1, $zero, 90
+	st.w $t1, $t0, 0
+# %op4 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -26
+# br i1 %op4, label %label5, label %label6
+	ld.b $t0, $fp, -26
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label5
+	b .main_label6
+.main_label5:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label6:
+# %op7 = getelementptr [1 x i32], [1 x i32]* @y, i32 0, i32 0
+	la.local $t0, y
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -34
+# store i32 18, i32* %op7
+	ld.d $t0, $fp, -34
+	addi.w $t1, $zero, 18
+	st.w $t1, $t0, 0
+# %op8 = getelementptr [1 x i32], [1 x i32]* @x, i32 0, i32 0
+	la.local $t0, x
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -42
+# %op9 = getelementptr [1 x i32], [1 x i32]* @y, i32 0, i32 0
+	la.local $t0, y
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -50
+# %op10 = call i32 @funArray(i32* %op8, i32* %op9)
+	ld.d $a0, $fp, -42
+	ld.d $a1, $fp, -50
+	bl funArray
+	st.w $a0, $fp, -54
+# ret i32 %op10
+	ld.w $a0, $fp, -54
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 64
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/21-comment.out

+0

tests/3-codegen/autogen/output/21-comment.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -16
+.main_label_entry:
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 16
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/3-decl_float.out

+0

tests/3-codegen/autogen/output/3-decl_float.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.main_label_entry:
+# %op0 = alloca float
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/3-output.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/3-output.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  call void @output(i32 11)
+  call void @output(i32 22222)
+  ret i32 0
+}

tests/3-codegen/autogen/output/3-output.out

+11
+22222
+0

tests/3-codegen/autogen/output/3-output.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -16
+.main_label_entry:
+# call void @output(i32 11)
+	addi.w $a0, $zero, 11
+	bl output
+# call void @output(i32 22222)
+	lu12i.w $a0, 5
+	ori $a0, $a0, 1742
+	bl output
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 16
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/4-decl_int_array.out

+0

tests/3-codegen/autogen/output/4-decl_int_array.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -64
+.main_label_entry:
+# %op0 = alloca [10 x i32]
+	addi.d $t0, $fp, -64
+	st.d $t0, $fp, -24
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 64
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/4-if.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/4-if.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  %op0 = alloca i32
+  %op1 = alloca i32
+  %op2 = alloca i32
+  store i32 11, i32* %op0
+  store i32 22, i32* %op1
+  store i32 33, i32* %op2
+  %op3 = load i32, i32* %op0
+  %op4 = load i32, i32* %op1
+  %op5 = icmp sgt i32 %op3, %op4
+  %op6 = zext i1 %op5 to i32
+  %op7 = icmp ne i32 %op6, 0
+  br i1 %op7, label %label8, label %label15
+label8:                                                ; preds = %label_entry
+  %op9 = load i32, i32* %op0
+  %op10 = load i32, i32* %op2
+  %op11 = icmp sgt i32 %op9, %op10
+  %op12 = zext i1 %op11 to i32
+  %op13 = icmp ne i32 %op12, 0
+  br i1 %op13, label %label21, label %label24
+label14:                                                ; preds = %label23, %label28
+  ret i32 0
+label15:                                                ; preds = %label_entry
+  %op16 = load i32, i32* %op2
+  %op17 = load i32, i32* %op1
+  %op18 = icmp slt i32 %op16, %op17
+  %op19 = zext i1 %op18 to i32
+  %op20 = icmp ne i32 %op19, 0
+  br i1 %op20, label %label26, label %label29
+label21:                                                ; preds = %label8
+  %op22 = load i32, i32* %op0
+  call void @output(i32 %op22)
+  br label %label23
+label23:                                                ; preds = %label21, %label24
+  br label %label14
+label24:                                                ; preds = %label8
+  %op25 = load i32, i32* %op2
+  call void @output(i32 %op25)
+  br label %label23
+label26:                                                ; preds = %label15
+  %op27 = load i32, i32* %op1
+  call void @output(i32 %op27)
+  br label %label28
+label28:                                                ; preds = %label26, %label29
+  br label %label14
+label29:                                                ; preds = %label15
+  %op30 = load i32, i32* %op2
+  call void @output(i32 %op30)
+  br label %label28
+}

tests/3-codegen/autogen/output/4-if.out

+33
+0

tests/3-codegen/autogen/output/4-if.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -112
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# %op2 = alloca i32
+	addi.d $t0, $fp, -52
+	st.d $t0, $fp, -48
+# store i32 11, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 11
+	st.w $t1, $t0, 0
+# store i32 22, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 22
+	st.w $t1, $t0, 0
+# store i32 33, i32* %op2
+	ld.d $t0, $fp, -48
+	addi.w $t1, $zero, 33
+	st.w $t1, $t0, 0
+# %op3 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op4 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -60
+# %op5 = icmp sgt i32 %op3, %op4
+	ld.w $t0, $fp, -56
+	ld.w $t1, $fp, -60
+	slt $t2, $t1, $t0
+	st.b $t2, $fp, -61
+# %op6 = zext i1 %op5 to i32
+	ld.b $t0, $fp, -61
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -65
+# %op7 = icmp ne i32 %op6, 0
+	ld.w $t0, $fp, -65
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -66
+# br i1 %op7, label %label8, label %label15
+	ld.b $t0, $fp, -66
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label8
+	b .main_label15
+.main_label8:
+# %op9 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -70
+# %op10 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -74
+# %op11 = icmp sgt i32 %op9, %op10
+	ld.w $t0, $fp, -70
+	ld.w $t1, $fp, -74
+	slt $t2, $t1, $t0
+	st.b $t2, $fp, -75
+# %op12 = zext i1 %op11 to i32
+	ld.b $t0, $fp, -75
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -79
+# %op13 = icmp ne i32 %op12, 0
+	ld.w $t0, $fp, -79
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -80
+# br i1 %op13, label %label21, label %label24
+	ld.b $t0, $fp, -80
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label21
+	b .main_label24
+.main_label14:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label15:
+# %op16 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -84
+# %op17 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -88
+# %op18 = icmp slt i32 %op16, %op17
+	ld.w $t0, $fp, -84
+	ld.w $t1, $fp, -88
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -89
+# %op19 = zext i1 %op18 to i32
+	ld.b $t0, $fp, -89
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -93
+# %op20 = icmp ne i32 %op19, 0
+	ld.w $t0, $fp, -93
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -94
+# br i1 %op20, label %label26, label %label29
+	ld.b $t0, $fp, -94
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label26
+	b .main_label29
+.main_label21:
+# %op22 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -98
+# call void @output(i32 %op22)
+	ld.w $a0, $fp, -98
+	bl output
+# br label %label23
+	b .main_label23
+.main_label23:
+# br label %label14
+	b .main_label14
+.main_label24:
+# %op25 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -102
+# call void @output(i32 %op25)
+	ld.w $a0, $fp, -102
+	bl output
+# br label %label23
+	b .main_label23
+.main_label26:
+# %op27 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -106
+# call void @output(i32 %op27)
+	ld.w $a0, $fp, -106
+	bl output
+# br label %label28
+	b .main_label28
+.main_label28:
+# br label %label14
+	b .main_label14
+.main_label29:
+# %op30 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -110
+# call void @output(i32 %op30)
+	ld.w $a0, $fp, -110
+	bl output
+# br label %label28
+	b .main_label28
+main_exit:
+	addi.d $sp, $sp, 112
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/5-decl_float_array.out

+0

tests/3-codegen/autogen/output/5-decl_float_array.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -64
+.main_label_entry:
+# %op0 = alloca [10 x float]
+	addi.d $t0, $fp, -64
+	st.d $t0, $fp, -24
+# ret void
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 64
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/5-while.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/5-while.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  %op0 = alloca i32
+  %op1 = alloca i32
+  store i32 10, i32* %op0
+  store i32 0, i32* %op1
+  br label %label2
+label2:                                                ; preds = %label_entry, %label8
+  %op3 = load i32, i32* %op1
+  %op4 = load i32, i32* %op0
+  %op5 = icmp slt i32 %op3, %op4
+  %op6 = zext i1 %op5 to i32
+  %op7 = icmp ne i32 %op6, 0
+  br i1 %op7, label %label8, label %label12
+label8:                                                ; preds = %label2
+  %op9 = load i32, i32* %op1
+  call void @output(i32 %op9)
+  %op10 = load i32, i32* %op1
+  %op11 = add i32 %op10, 1
+  store i32 %op11, i32* %op1
+  br label %label2
+label12:                                                ; preds = %label2
+  ret i32 0
+}

tests/3-codegen/autogen/output/5-while.out

+0
+1
+2
+3
+4
+5
+6
+7
+8
+9
+0

tests/3-codegen/autogen/output/5-while.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# store i32 10, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# store i32 0, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# br label %label2
+	b .main_label2
+.main_label2:
+# %op3 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -44
+# %op4 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -48
+# %op5 = icmp slt i32 %op3, %op4
+	ld.w $t0, $fp, -44
+	ld.w $t1, $fp, -48
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -49
+# %op6 = zext i1 %op5 to i32
+	ld.b $t0, $fp, -49
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -53
+# %op7 = icmp ne i32 %op6, 0
+	ld.w $t0, $fp, -53
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -54
+# br i1 %op7, label %label8, label %label12
+	ld.b $t0, $fp, -54
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label8
+	b .main_label12
+.main_label8:
+# %op9 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -58
+# call void @output(i32 %op9)
+	ld.w $a0, $fp, -58
+	bl output
+# %op10 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -62
+# %op11 = add i32 %op10, 1
+	ld.w $t0, $fp, -62
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -66
+# store i32 %op11, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t1, $fp, -66
+	st.w $t1, $t0, 0
+# br label %label2
+	b .main_label2
+.main_label12:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/6-array.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/6-array.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @main() {
+label_entry:
+  %op0 = alloca [10 x i32]
+  %op1 = alloca i32
+  store i32 0, i32* %op1
+  %op2 = icmp slt i32 0, 0
+  br i1 %op2, label %label3, label %label4
+label3:                                                ; preds = %label_entry
+  call void @neg_idx_except()
+  ret i32 0
+label4:                                                ; preds = %label_entry
+  %op5 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0
+  store i32 11, i32* %op5
+  %op6 = icmp slt i32 4, 0
+  br i1 %op6, label %label7, label %label8
+label7:                                                ; preds = %label4
+  call void @neg_idx_except()
+  ret i32 0
+label8:                                                ; preds = %label4
+  %op9 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 4
+  store i32 22, i32* %op9
+  %op10 = icmp slt i32 9, 0
+  br i1 %op10, label %label11, label %label12
+label11:                                                ; preds = %label8
+  call void @neg_idx_except()
+  ret i32 0
+label12:                                                ; preds = %label8
+  %op13 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 9
+  store i32 33, i32* %op13
+  %op14 = icmp slt i32 0, 0
+  br i1 %op14, label %label15, label %label16
+label15:                                                ; preds = %label12
+  call void @neg_idx_except()
+  ret i32 0
+label16:                                                ; preds = %label12
+  %op17 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0
+  %op18 = load i32, i32* %op17
+  call void @output(i32 %op18)
+  %op19 = icmp slt i32 4, 0
+  br i1 %op19, label %label20, label %label21
+label20:                                                ; preds = %label16
+  call void @neg_idx_except()
+  ret i32 0
+label21:                                                ; preds = %label16
+  %op22 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 4
+  %op23 = load i32, i32* %op22
+  call void @output(i32 %op23)
+  %op24 = icmp slt i32 9, 0
+  br i1 %op24, label %label25, label %label26
+label25:                                                ; preds = %label21
+  call void @neg_idx_except()
+  ret i32 0
+label26:                                                ; preds = %label21
+  %op27 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 9
+  %op28 = load i32, i32* %op27
+  call void @output(i32 %op28)
+  ret i32 0
+}

tests/3-codegen/autogen/output/6-array.out

+11
+22
+33
+0

tests/3-codegen/autogen/output/6-array.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -144
+.main_label_entry:
+# %op0 = alloca [10 x i32]
+	addi.d $t0, $fp, -64
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -76
+	st.d $t0, $fp, -72
+# store i32 0, i32* %op1
+	ld.d $t0, $fp, -72
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# %op2 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -77
+# br i1 %op2, label %label3, label %label4
+	ld.b $t0, $fp, -77
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label3
+	b .main_label4
+.main_label3:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label4:
+# %op5 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -85
+# store i32 11, i32* %op5
+	ld.d $t0, $fp, -85
+	addi.w $t1, $zero, 11
+	st.w $t1, $t0, 0
+# %op6 = icmp slt i32 4, 0
+	addi.w $t0, $zero, 4
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -86
+# br i1 %op6, label %label7, label %label8
+	ld.b $t0, $fp, -86
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label7
+	b .main_label8
+.main_label7:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label8:
+# %op9 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 4
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 4
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -94
+# store i32 22, i32* %op9
+	ld.d $t0, $fp, -94
+	addi.w $t1, $zero, 22
+	st.w $t1, $t0, 0
+# %op10 = icmp slt i32 9, 0
+	addi.w $t0, $zero, 9
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -95
+# br i1 %op10, label %label11, label %label12
+	ld.b $t0, $fp, -95
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label11
+	b .main_label12
+.main_label11:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label12:
+# %op13 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 9
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 9
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -103
+# store i32 33, i32* %op13
+	ld.d $t0, $fp, -103
+	addi.w $t1, $zero, 33
+	st.w $t1, $t0, 0
+# %op14 = icmp slt i32 0, 0
+	addi.w $t0, $zero, 0
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -104
+# br i1 %op14, label %label15, label %label16
+	ld.b $t0, $fp, -104
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label15
+	b .main_label16
+.main_label15:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label16:
+# %op17 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -112
+# %op18 = load i32, i32* %op17
+	ld.d $t0, $fp, -112
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -116
+# call void @output(i32 %op18)
+	ld.w $a0, $fp, -116
+	bl output
+# %op19 = icmp slt i32 4, 0
+	addi.w $t0, $zero, 4
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -117
+# br i1 %op19, label %label20, label %label21
+	ld.b $t0, $fp, -117
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label20
+	b .main_label21
+.main_label20:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label21:
+# %op22 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 4
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 4
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -125
+# %op23 = load i32, i32* %op22
+	ld.d $t0, $fp, -125
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -129
+# call void @output(i32 %op23)
+	ld.w $a0, $fp, -129
+	bl output
+# %op24 = icmp slt i32 9, 0
+	addi.w $t0, $zero, 9
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -130
+# br i1 %op24, label %label25, label %label26
+	ld.b $t0, $fp, -130
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label25
+	b .main_label26
+.main_label25:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label26:
+# %op27 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 9
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 9
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -138
+# %op28 = load i32, i32* %op27
+	ld.d $t0, $fp, -138
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -142
+# call void @output(i32 %op28)
+	ld.w $a0, $fp, -142
+	bl output
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 144
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/6-num_add_int.out

+210

tests/3-codegen/autogen/output/6-num_add_int.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -48
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = add i32 1000, 234
+	addi.w $t0, $zero, 1000
+	addi.w $t1, $zero, 234
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -32
+# store i32 %op1, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -32
+	st.w $t1, $t0, 0
+# %op2 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -36
+# ret i32 %op2
+	ld.w $a0, $fp, -36
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 48
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/7-assign_int_var_local.out

+210

tests/3-codegen/autogen/output/7-assign_int_var_local.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -32
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# store i32 1234, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 1234
+	st.w $t1, $t0, 0
+# %op1 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -32
+# ret i32 %op1
+	ld.w $a0, $fp, -32
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 32
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/7-function.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/7-function.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @min(i32 %arg0, i32 %arg1) {
+label_entry:
+  %op2 = alloca i32
+  store i32 %arg0, i32* %op2
+  %op3 = alloca i32
+  store i32 %arg1, i32* %op3
+  %op4 = load i32, i32* %op2
+  %op5 = load i32, i32* %op3
+  %op6 = icmp sle i32 %op4, %op5
+  %op7 = zext i1 %op6 to i32
+  %op8 = icmp ne i32 %op7, 0
+  br i1 %op8, label %label9, label %label12
+label9:                                                ; preds = %label_entry
+  %op10 = load i32, i32* %op2
+  ret i32 %op10
+label11:
+  ret i32 0
+label12:                                                ; preds = %label_entry
+  %op13 = load i32, i32* %op3
+  ret i32 %op13
+}
+define i32 @main() {
+label_entry:
+  %op0 = alloca i32
+  %op1 = alloca i32
+  %op2 = alloca i32
+  store i32 11, i32* %op0
+  store i32 22, i32* %op1
+  store i32 33, i32* %op2
+  %op3 = load i32, i32* %op0
+  %op4 = load i32, i32* %op1
+  %op5 = call i32 @min(i32 %op3, i32 %op4)
+  call void @output(i32 %op5)
+  %op6 = load i32, i32* %op1
+  %op7 = load i32, i32* %op2
+  %op8 = call i32 @min(i32 %op6, i32 %op7)
+  call void @output(i32 %op8)
+  %op9 = load i32, i32* %op2
+  %op10 = load i32, i32* %op0
+  %op11 = call i32 @min(i32 %op9, i32 %op10)
+  call void @output(i32 %op11)
+  ret i32 0
+}

tests/3-codegen/autogen/output/7-function.out

+11
+22
+11
+0

tests/3-codegen/autogen/output/7-function.s

+	.text
+	.globl min
+	.type min, @function
+min:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+	st.w $a0, $fp, -20
+	st.w $a1, $fp, -24
+.min_label_entry:
+# %op2 = alloca i32
+	addi.d $t0, $fp, -36
+	st.d $t0, $fp, -32
+# store i32 %arg0, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# %op3 = alloca i32
+	addi.d $t0, $fp, -48
+	st.d $t0, $fp, -44
+# store i32 %arg1, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t1, $fp, -24
+	st.w $t1, $t0, 0
+# %op4 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -52
+# %op5 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op6 = icmp sle i32 %op4, %op5
+	ld.w $t0, $fp, -52
+	ld.w $t1, $fp, -56
+	slt $t2, $t1, $t0
+	xori $t2, $t2, 1
+	st.b $t2, $fp, -57
+# %op7 = zext i1 %op6 to i32
+	ld.b $t0, $fp, -57
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -61
+# %op8 = icmp ne i32 %op7, 0
+	ld.w $t0, $fp, -61
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -62
+# br i1 %op8, label %label9, label %label12
+	ld.b $t0, $fp, -62
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .min_label9
+	b .min_label12
+.min_label9:
+# %op10 = load i32, i32* %op2
+	ld.d $t0, $fp, -32
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -66
+# ret i32 %op10
+	ld.w $a0, $fp, -66
+	b min_exit
+.min_label11:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b min_exit
+.min_label12:
+# %op13 = load i32, i32* %op3
+	ld.d $t0, $fp, -44
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -70
+# ret i32 %op13
+	ld.w $a0, $fp, -70
+	b min_exit
+min_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -96
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# %op2 = alloca i32
+	addi.d $t0, $fp, -52
+	st.d $t0, $fp, -48
+# store i32 11, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 11
+	st.w $t1, $t0, 0
+# store i32 22, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 22
+	st.w $t1, $t0, 0
+# store i32 33, i32* %op2
+	ld.d $t0, $fp, -48
+	addi.w $t1, $zero, 33
+	st.w $t1, $t0, 0
+# %op3 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op4 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -60
+# %op5 = call i32 @min(i32 %op3, i32 %op4)
+	ld.w $a0, $fp, -56
+	ld.w $a1, $fp, -60
+	bl min
+	st.w $a0, $fp, -64
+# call void @output(i32 %op5)
+	ld.w $a0, $fp, -64
+	bl output
+# %op6 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -68
+# %op7 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -72
+# %op8 = call i32 @min(i32 %op6, i32 %op7)
+	ld.w $a0, $fp, -68
+	ld.w $a1, $fp, -72
+	bl min
+	st.w $a0, $fp, -76
+# call void @output(i32 %op8)
+	ld.w $a0, $fp, -76
+	bl output
+# %op9 = load i32, i32* %op2
+	ld.d $t0, $fp, -48
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -80
+# %op10 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -84
+# %op11 = call i32 @min(i32 %op9, i32 %op10)
+	ld.w $a0, $fp, -80
+	ld.w $a1, $fp, -84
+	bl min
+	st.w $a0, $fp, -88
+# call void @output(i32 %op11)
+	ld.w $a0, $fp, -88
+	bl output
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 96
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/8-assign_int_array_local.out

+210

tests/3-codegen/autogen/output/8-assign_int_array_local.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -96
+.main_label_entry:
+# %op0 = alloca [10 x i32]
+	addi.d $t0, $fp, -64
+	st.d $t0, $fp, -24
+# %op1 = icmp slt i32 3, 0
+	addi.w $t0, $zero, 3
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -65
+# br i1 %op1, label %label2, label %label3
+	ld.b $t0, $fp, -65
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label2
+	b .main_label3
+.main_label2:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label3:
+# %op4 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 3
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 3
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -73
+# store i32 1234, i32* %op4
+	ld.d $t0, $fp, -73
+	addi.w $t1, $zero, 1234
+	st.w $t1, $t0, 0
+# %op5 = icmp slt i32 3, 0
+	addi.w $t0, $zero, 3
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -74
+# br i1 %op5, label %label6, label %label7
+	ld.b $t0, $fp, -74
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label6
+	b .main_label7
+.main_label6:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label7:
+# %op8 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 3
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 3
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -82
+# %op9 = load i32, i32* %op8
+	ld.d $t0, $fp, -82
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -86
+# ret i32 %op9
+	ld.w $a0, $fp, -86
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 96
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/8-store.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/8-store.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @store(i32* %arg0, i32 %arg1, i32 %arg2) {
+label_entry:
+  %op3 = alloca i32*
+  store i32* %arg0, i32** %op3
+  %op4 = alloca i32
+  store i32 %arg1, i32* %op4
+  %op5 = alloca i32
+  store i32 %arg2, i32* %op5
+  %op6 = load i32, i32* %op5
+  %op7 = load i32, i32* %op4
+  %op8 = icmp slt i32 %op7, 0
+  br i1 %op8, label %label9, label %label10
+label9:                                                ; preds = %label_entry
+  call void @neg_idx_except()
+  ret i32 0
+label10:                                                ; preds = %label_entry
+  %op11 = load i32*, i32** %op3
+  %op12 = getelementptr i32, i32* %op11, i32 %op7
+  store i32 %op6, i32* %op12
+  %op13 = load i32, i32* %op5
+  ret i32 %op13
+}
+define i32 @main() {
+label_entry:
+  %op0 = alloca [10 x i32]
+  %op1 = alloca i32
+  %op2 = alloca i32
+  store i32 0, i32* %op1
+  br label %label3
+label3:                                                ; preds = %label_entry, %label8
+  %op4 = load i32, i32* %op1
+  %op5 = icmp slt i32 %op4, 10
+  %op6 = zext i1 %op5 to i32
+  %op7 = icmp ne i32 %op6, 0
+  br i1 %op7, label %label8, label %label16
+label8:                                                ; preds = %label3
+  %op9 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0
+  %op10 = load i32, i32* %op1
+  %op11 = load i32, i32* %op1
+  %op12 = mul i32 %op11, 2
+  %op13 = call i32 @store(i32* %op9, i32 %op10, i32 %op12)
+  %op14 = load i32, i32* %op1
+  %op15 = add i32 %op14, 1
+  store i32 %op15, i32* %op1
+  br label %label3
+label16:                                                ; preds = %label3
+  store i32 0, i32* %op2
+  store i32 0, i32* %op1
+  br label %label17
+label17:                                                ; preds = %label16, %label29
+  %op18 = load i32, i32* %op1
+  %op19 = icmp slt i32 %op18, 10
+  %op20 = zext i1 %op19 to i32
+  %op21 = icmp ne i32 %op20, 0
+  br i1 %op21, label %label22, label %label26
+label22:                                                ; preds = %label17
+  %op23 = load i32, i32* %op2
+  %op24 = load i32, i32* %op1
+  %op25 = icmp slt i32 %op24, 0
+  br i1 %op25, label %label28, label %label29
+label26:                                                ; preds = %label17
+  %op27 = load i32, i32* %op2
+  call void @output(i32 %op27)
+  ret i32 0
+label28:                                                ; preds = %label22
+  call void @neg_idx_except()
+  ret i32 0
+label29:                                                ; preds = %label22
+  %op30 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 %op24
+  %op31 = load i32, i32* %op30
+  %op32 = add i32 %op23, %op31
+  store i32 %op32, i32* %op2
+  %op33 = load i32, i32* %op1
+  %op34 = add i32 %op33, 1
+  store i32 %op34, i32* %op1
+  br label %label17
+}

tests/3-codegen/autogen/output/8-store.out

+90
+0

tests/3-codegen/autogen/output/8-store.s

+	.text
+	.globl store
+	.type store, @function
+store:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -112
+	st.d $a0, $fp, -24
+	st.w $a1, $fp, -28
+	st.w $a2, $fp, -32
+.store_label_entry:
+# %op3 = alloca i32*
+	addi.d $t0, $fp, -48
+	st.d $t0, $fp, -40
+# store i32* %arg0, i32** %op3
+	ld.d $t0, $fp, -40
+	ld.d $t1, $fp, -24
+	st.d $t1, $t0, 0
+# %op4 = alloca i32
+	addi.d $t0, $fp, -60
+	st.d $t0, $fp, -56
+# store i32 %arg1, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t1, $fp, -28
+	st.w $t1, $t0, 0
+# %op5 = alloca i32
+	addi.d $t0, $fp, -72
+	st.d $t0, $fp, -68
+# store i32 %arg2, i32* %op5
+	ld.d $t0, $fp, -68
+	ld.w $t1, $fp, -32
+	st.w $t1, $t0, 0
+# %op6 = load i32, i32* %op5
+	ld.d $t0, $fp, -68
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -76
+# %op7 = load i32, i32* %op4
+	ld.d $t0, $fp, -56
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -80
+# %op8 = icmp slt i32 %op7, 0
+	ld.w $t0, $fp, -80
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -81
+# br i1 %op8, label %label9, label %label10
+	ld.b $t0, $fp, -81
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .store_label9
+	b .store_label10
+.store_label9:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b store_exit
+.store_label10:
+# %op11 = load i32*, i32** %op3
+	ld.d $t0, $fp, -40
+	ld.d $t0, $t0, 0
+	st.d $t0, $fp, -89
+# %op12 = getelementptr i32, i32* %op11, i32 %op7
+	ld.d $t0, $fp, -89
+	ld.w $t1, $fp, -80
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -97
+# store i32 %op6, i32* %op12
+	ld.d $t0, $fp, -97
+	ld.w $t1, $fp, -76
+	st.w $t1, $t0, 0
+# %op13 = load i32, i32* %op5
+	ld.d $t0, $fp, -68
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -101
+# ret i32 %op13
+	ld.w $a0, $fp, -101
+	b store_exit
+store_exit:
+	addi.d $sp, $sp, 112
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -192
+.main_label_entry:
+# %op0 = alloca [10 x i32]
+	addi.d $t0, $fp, -64
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -76
+	st.d $t0, $fp, -72
+# %op2 = alloca i32
+	addi.d $t0, $fp, -88
+	st.d $t0, $fp, -84
+# store i32 0, i32* %op1
+	ld.d $t0, $fp, -72
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# br label %label3
+	b .main_label3
+.main_label3:
+# %op4 = load i32, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -92
+# %op5 = icmp slt i32 %op4, 10
+	ld.w $t0, $fp, -92
+	addi.w $t1, $zero, 10
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -93
+# %op6 = zext i1 %op5 to i32
+	ld.b $t0, $fp, -93
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -97
+# %op7 = icmp ne i32 %op6, 0
+	ld.w $t0, $fp, -97
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -98
+# br i1 %op7, label %label8, label %label16
+	ld.b $t0, $fp, -98
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label8
+	b .main_label16
+.main_label8:
+# %op9 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -106
+# %op10 = load i32, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -110
+# %op11 = load i32, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -114
+# %op12 = mul i32 %op11, 2
+	ld.w $t0, $fp, -114
+	addi.w $t1, $zero, 2
+	mul.w $t2, $t0, $t1
+	st.w $t2, $fp, -118
+# %op13 = call i32 @store(i32* %op9, i32 %op10, i32 %op12)
+	ld.d $a0, $fp, -106
+	ld.w $a1, $fp, -110
+	ld.w $a2, $fp, -118
+	bl store
+	st.w $a0, $fp, -122
+# %op14 = load i32, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -126
+# %op15 = add i32 %op14, 1
+	ld.w $t0, $fp, -126
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -130
+# store i32 %op15, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t1, $fp, -130
+	st.w $t1, $t0, 0
+# br label %label3
+	b .main_label3
+.main_label16:
+# store i32 0, i32* %op2
+	ld.d $t0, $fp, -84
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# store i32 0, i32* %op1
+	ld.d $t0, $fp, -72
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# br label %label17
+	b .main_label17
+.main_label17:
+# %op18 = load i32, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -134
+# %op19 = icmp slt i32 %op18, 10
+	ld.w $t0, $fp, -134
+	addi.w $t1, $zero, 10
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -135
+# %op20 = zext i1 %op19 to i32
+	ld.b $t0, $fp, -135
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -139
+# %op21 = icmp ne i32 %op20, 0
+	ld.w $t0, $fp, -139
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -140
+# br i1 %op21, label %label22, label %label26
+	ld.b $t0, $fp, -140
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label22
+	b .main_label26
+.main_label22:
+# %op23 = load i32, i32* %op2
+	ld.d $t0, $fp, -84
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -144
+# %op24 = load i32, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -148
+# %op25 = icmp slt i32 %op24, 0
+	ld.w $t0, $fp, -148
+	addi.w $t1, $zero, 0
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -149
+# br i1 %op25, label %label28, label %label29
+	ld.b $t0, $fp, -149
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label28
+	b .main_label29
+.main_label26:
+# %op27 = load i32, i32* %op2
+	ld.d $t0, $fp, -84
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -153
+# call void @output(i32 %op27)
+	ld.w $a0, $fp, -153
+	bl output
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label28:
+# call void @neg_idx_except()
+	bl neg_idx_except
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+.main_label29:
+# %op30 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 %op24
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 0
+	addi.d $t2, $zero, 40
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	ld.w $t1, $fp, -148
+	addi.d $t2, $zero, 4
+	mul.d $t1, $t1, $t2
+	add.d $t0, $t0, $t1
+	st.d $t0, $fp, -161
+# %op31 = load i32, i32* %op30
+	ld.d $t0, $fp, -161
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -165
+# %op32 = add i32 %op23, %op31
+	ld.w $t0, $fp, -144
+	ld.w $t1, $fp, -165
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -169
+# store i32 %op32, i32* %op2
+	ld.d $t0, $fp, -84
+	ld.w $t1, $fp, -169
+	st.w $t1, $t0, 0
+# %op33 = load i32, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -173
+# %op34 = add i32 %op33, 1
+	ld.w $t0, $fp, -173
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -177
+# store i32 %op34, i32* %op1
+	ld.d $t0, $fp, -72
+	ld.w $t1, $fp, -177
+	st.w $t1, $t0, 0
+# br label %label17
+	b .main_label17
+main_exit:
+	addi.d $sp, $sp, 192
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/9-assign_cast.out

+5

tests/3-codegen/autogen/output/9-assign_cast.s

+	.text
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# %op2 = icmp slt i32 1, 3
+	addi.w $t0, $zero, 1
+	addi.w $t1, $zero, 3
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -41
+# %op3 = zext i1 %op2 to i32
+	ld.b $t0, $fp, -41
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -45
+# store i32 %op3, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t1, $fp, -45
+	st.w $t1, $t0, 0
+# %op4 = sitofp i32 2 to float
+	addi.w $t0, $zero, 2
+	movgr2fr.w $ft0, $t0
+	ffint.s.w $ft0, $ft0
+	fst.s $ft0, $fp, -49
+# %op5 = fadd float %op4, 0x4003333340000000
+	fld.s $ft0, $fp, -49
+	lu12i.w $t8, 262553
+	ori $t8, $t8, 2458
+	movgr2fr.w $ft1, $t8
+	fadd.s $ft2, $ft0, $ft1
+	fst.s $ft2, $fp, -53
+# %op6 = fptosi float %op5 to i32
+	fld.s $ft0, $fp, -53
+	ftintrz.w.s $ft0, $ft0
+	movfr2gr.s $t0, $ft0
+	st.w $t0, $fp, -57
+# store i32 %op6, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t1, $fp, -57
+	st.w $t1, $t0, 0
+# %op7 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -61
+# %op8 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -65
+# %op9 = add i32 %op7, %op8
+	ld.w $t0, $fp, -61
+	ld.w $t1, $fp, -65
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -69
+# ret i32 %op9
+	ld.w $a0, $fp, -69
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/output/9-fibonacci.ll

+; ModuleID = 'cminus'
+source_filename = "/code/compiler/24.ta/tests/3-codegen/autogen/testcases/9-fibonacci.cminus"
+
+declare i32 @input()
+
+declare void @output(i32)
+
+declare void @outputFloat(float)
+
+declare void @neg_idx_except()
+
+define i32 @fibonacci(i32 %arg0) {
+label_entry:
+  %op1 = alloca i32
+  store i32 %arg0, i32* %op1
+  %op2 = load i32, i32* %op1
+  %op3 = icmp eq i32 %op2, 0
+  %op4 = zext i1 %op3 to i32
+  %op5 = icmp ne i32 %op4, 0
+  br i1 %op5, label %label6, label %label8
+label6:                                                ; preds = %label_entry
+  ret i32 0
+label7:                                                ; preds = %label14
+  ret i32 0
+label8:                                                ; preds = %label_entry
+  %op9 = load i32, i32* %op1
+  %op10 = icmp eq i32 %op9, 1
+  %op11 = zext i1 %op10 to i32
+  %op12 = icmp ne i32 %op11, 0
+  br i1 %op12, label %label13, label %label15
+label13:                                                ; preds = %label8
+  ret i32 1
+label14:
+  br label %label7
+label15:                                                ; preds = %label8
+  %op16 = load i32, i32* %op1
+  %op17 = sub i32 %op16, 1
+  %op18 = call i32 @fibonacci(i32 %op17)
+  %op19 = load i32, i32* %op1
+  %op20 = sub i32 %op19, 2
+  %op21 = call i32 @fibonacci(i32 %op20)
+  %op22 = add i32 %op18, %op21
+  ret i32 %op22
+}
+define i32 @main() {
+label_entry:
+  %op0 = alloca i32
+  %op1 = alloca i32
+  store i32 10, i32* %op0
+  store i32 0, i32* %op1
+  br label %label2
+label2:                                                ; preds = %label_entry, %label8
+  %op3 = load i32, i32* %op1
+  %op4 = load i32, i32* %op0
+  %op5 = icmp slt i32 %op3, %op4
+  %op6 = zext i1 %op5 to i32
+  %op7 = icmp ne i32 %op6, 0
+  br i1 %op7, label %label8, label %label13
+label8:                                                ; preds = %label2
+  %op9 = load i32, i32* %op1
+  %op10 = call i32 @fibonacci(i32 %op9)
+  call void @output(i32 %op10)
+  %op11 = load i32, i32* %op1
+  %op12 = add i32 %op11, 1
+  store i32 %op12, i32* %op1
+  br label %label2
+label13:                                                ; preds = %label2
+  ret i32 0
+}

tests/3-codegen/autogen/output/9-fibonacci.out

+0
+1
+1
+2
+3
+5
+8
+13
+21
+34
+0

tests/3-codegen/autogen/output/9-fibonacci.s

+	.text
+	.globl fibonacci
+	.type fibonacci, @function
+fibonacci:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+	st.w $a0, $fp, -20
+.fibonacci_label_entry:
+# %op1 = alloca i32
+	addi.d $t0, $fp, -32
+	st.d $t0, $fp, -28
+# store i32 %arg0, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t1, $fp, -20
+	st.w $t1, $t0, 0
+# %op2 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -36
+# %op3 = icmp eq i32 %op2, 0
+	ld.w $t0, $fp, -36
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltui $t2, $t2, 1
+	st.b $t2, $fp, -37
+# %op4 = zext i1 %op3 to i32
+	ld.b $t0, $fp, -37
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -41
+# %op5 = icmp ne i32 %op4, 0
+	ld.w $t0, $fp, -41
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -42
+# br i1 %op5, label %label6, label %label8
+	ld.b $t0, $fp, -42
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .fibonacci_label6
+	b .fibonacci_label8
+.fibonacci_label6:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b fibonacci_exit
+.fibonacci_label7:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b fibonacci_exit
+.fibonacci_label8:
+# %op9 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -46
+# %op10 = icmp eq i32 %op9, 1
+	ld.w $t0, $fp, -46
+	addi.w $t1, $zero, 1
+	xor $t2, $t0, $t1
+	sltui $t2, $t2, 1
+	st.b $t2, $fp, -47
+# %op11 = zext i1 %op10 to i32
+	ld.b $t0, $fp, -47
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -51
+# %op12 = icmp ne i32 %op11, 0
+	ld.w $t0, $fp, -51
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -52
+# br i1 %op12, label %label13, label %label15
+	ld.b $t0, $fp, -52
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .fibonacci_label13
+	b .fibonacci_label15
+.fibonacci_label13:
+# ret i32 1
+	addi.w $a0, $zero, 1
+	b fibonacci_exit
+.fibonacci_label14:
+# br label %label7
+	b .fibonacci_label7
+.fibonacci_label15:
+# %op16 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -56
+# %op17 = sub i32 %op16, 1
+	ld.w $t0, $fp, -56
+	addi.w $t1, $zero, 1
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -60
+# %op18 = call i32 @fibonacci(i32 %op17)
+	ld.w $a0, $fp, -60
+	bl fibonacci
+	st.w $a0, $fp, -64
+# %op19 = load i32, i32* %op1
+	ld.d $t0, $fp, -28
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -68
+# %op20 = sub i32 %op19, 2
+	ld.w $t0, $fp, -68
+	addi.w $t1, $zero, 2
+	sub.w $t2, $t0, $t1
+	st.w $t2, $fp, -72
+# %op21 = call i32 @fibonacci(i32 %op20)
+	ld.w $a0, $fp, -72
+	bl fibonacci
+	st.w $a0, $fp, -76
+# %op22 = add i32 %op18, %op21
+	ld.w $t0, $fp, -64
+	ld.w $t1, $fp, -76
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -80
+# ret i32 %op22
+	ld.w $a0, $fp, -80
+	b fibonacci_exit
+fibonacci_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra
+	.globl main
+	.type main, @function
+main:
+	st.d $ra, $sp, -8
+	st.d $fp, $sp, -16
+	addi.d $fp, $sp, 0
+	addi.d $sp, $sp, -80
+.main_label_entry:
+# %op0 = alloca i32
+	addi.d $t0, $fp, -28
+	st.d $t0, $fp, -24
+# %op1 = alloca i32
+	addi.d $t0, $fp, -40
+	st.d $t0, $fp, -36
+# store i32 10, i32* %op0
+	ld.d $t0, $fp, -24
+	addi.w $t1, $zero, 10
+	st.w $t1, $t0, 0
+# store i32 0, i32* %op1
+	ld.d $t0, $fp, -36
+	addi.w $t1, $zero, 0
+	st.w $t1, $t0, 0
+# br label %label2
+	b .main_label2
+.main_label2:
+# %op3 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -44
+# %op4 = load i32, i32* %op0
+	ld.d $t0, $fp, -24
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -48
+# %op5 = icmp slt i32 %op3, %op4
+	ld.w $t0, $fp, -44
+	ld.w $t1, $fp, -48
+	slt $t2, $t0, $t1
+	st.b $t2, $fp, -49
+# %op6 = zext i1 %op5 to i32
+	ld.b $t0, $fp, -49
+	bstrpick.w $t0, $t0, 0, 0
+	st.w $t0, $fp, -53
+# %op7 = icmp ne i32 %op6, 0
+	ld.w $t0, $fp, -53
+	addi.w $t1, $zero, 0
+	xor $t2, $t0, $t1
+	sltu $t2, $zero, $t2
+	st.b $t2, $fp, -54
+# br i1 %op7, label %label8, label %label13
+	ld.b $t0, $fp, -54
+	bstrpick.d $t0, $t0, 0, 0
+	bne $t0, $zero, .main_label8
+	b .main_label13
+.main_label8:
+# %op9 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -58
+# %op10 = call i32 @fibonacci(i32 %op9)
+	ld.w $a0, $fp, -58
+	bl fibonacci
+	st.w $a0, $fp, -62
+# call void @output(i32 %op10)
+	ld.w $a0, $fp, -62
+	bl output
+# %op11 = load i32, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t0, $t0, 0
+	st.w $t0, $fp, -66
+# %op12 = add i32 %op11, 1
+	ld.w $t0, $fp, -66
+	addi.w $t1, $zero, 1
+	add.w $t2, $t0, $t1
+	st.w $t2, $fp, -70
+# store i32 %op12, i32* %op1
+	ld.d $t0, $fp, -36
+	ld.w $t1, $fp, -70
+	st.w $t1, $t0, 0
+# br label %label2
+	b .main_label2
+.main_label13:
+# ret i32 0
+	addi.w $a0, $zero, 0
+	b main_exit
+main_exit:
+	addi.d $sp, $sp, 80
+	ld.d $ra, $sp, -8
+	ld.d $fp, $sp, -16
+	jr $ra

tests/3-codegen/autogen/testcases/0-io.cminus

+int main(void) {
+    output(input());
+    return 0;
+}

tests/3-codegen/autogen/testcases/0-io.in

+1234

tests/3-codegen/autogen/testcases/0-io.out

+1234
+0

tests/3-codegen/autogen/testcases/1-return.cminus

+int main(void) {
+    output(111);
+    return 111;
+}

tests/3-codegen/autogen/testcases/1-return.out

+111
+111

tests/3-codegen/autogen/testcases/1-return.s

+program
+--fun-declaration: main
+----compound-stmt
+------expression-stmt
+--------simple-expression
+----------additive-expression
+------------term
+--------------call: output()
+----------------simple-expression
+------------------additive-expression
+--------------------term
+----------------------num (int): 111
+------return-stmt
+--------simple-expression
+----------additive-expression
+------------term
+--------------num (int): 111

tests/3-codegen/autogen/testcases/10-float.cminus

+int main(void) {
+    float a;
+    float b;
+    float c;
+
+    a = 1.1;
+    b = 1.5;
+    c = 1.2;
+
+    outputFloat(a * b + c);
+    return 0;
+}
+
+/*
+    用 gcc 编译此文件生成汇编代码时，命令为  gcc -include io.h -S 10-float.c
+    其中 io.h 位于 src/io/io.h，
+
+    也可以在本文件开头加上    void outputFloat(float x);
+*/

tests/3-codegen/autogen/testcases/10-float.out

+2.850000
+0

tests/3-codegen/autogen/testcases/11-floatcall.cminus

+/* float function call */
+
+float mod(float x, float y) {
+    int div;
+    div = x / y;
+    return x - div * y;
+}
+
+int main(void) {
+    float a;
+    float b;
+
+    a = 11.2;
+    b = 2.2;
+
+    outputFloat(mod(a, b));
+    return 0;
+}

tests/3-codegen/autogen/testcases/11-floatcall.out

+0.200000
+0

tests/3-codegen/autogen/testcases/12-global.cminus

+int seed;
+
+int randomLCG(void) {
+    seed = seed * 1103515245 + 12345;
+    return seed;
+}
+
+int randBin(void) {
+    if (randomLCG() > 0)
+        return 1;
+    else
+        return 0;
+}
+
+/* random walk */
+int returnToZeroSteps(void) {
+    int x;
+    int steps;
+
+    x = 0;
+    steps = 0;
+
+    while (steps < 20) {
+        if (randBin())
+            x = x + 1;
+        else
+            x = x - 1;
+
+        steps = steps + 1;
+        if (x == 0)
+            return steps;
+    }
+    return 20;
+}
+
+int main(void) {
+    int i;
+    i = 0;
+
+    seed = 3407;
+
+    while (i < 20) {
+        output(returnToZeroSteps());
+        i = i + 1;
+    }
+    return 0;
+}

tests/3-codegen/autogen/testcases/12-global.out

+4
+2
+2
+4
+8
+2
+2
+2
+2
+2
+6
+2
+10
+8
+4
+2
+20
+2
+2
+8
+0

tests/3-codegen/autogen/testcases/13-complex.cminus

+/* 01 背包问题 */
+
+int n;
+int m;
+
+int w[5];
+int v[5];
+
+int dp[66]; /* dp[n * 11 + size] 表示前 n 个物品放入容量为 size 的背包中的最大价值，初始化为 -1 */
+
+int max(int a, int b) {
+    if (a > b)
+        return a;
+    else
+        return b;
+}
+
+/* 状态转移方程：
+   dp[n][size] = max(dp[n - 1][size], dp[n - 1][size - w[n]] + v[n])
+   边界条件：
+   dp[n][size <= 0] = 0
+   dp[0][size] = 0  */
+
+int knapsack(int n, int size) {
+    int result;
+    if (size <= 0)
+        return 0;
+    if (n == 0)
+        return 0;
+    if (dp[n * 11 + size] >= 0)
+        return dp[n * 11 + size];
+
+    if (size < w[n - 1])
+        result = knapsack(n - 1, size);
+    else
+        result = max(knapsack(n - 1, size), knapsack(n - 1, size - w[n - 1]) + v[n - 1]);
+
+    dp[n * 11 + size] = result;
+    return result;
+}
+
+int main(void) {
+    int i;
+    i = 0;
+
+    n = 5;
+    m = 10;
+    w[0] = 2;
+    w[1] = 2;
+    w[2] = 6;
+    w[3] = 5;
+    w[4] = 4;
+
+    v[0] = 6;
+    v[1] = 3;
+    v[2] = 5;
+    v[3] = 4;
+    v[4] = 6;
+
+    while (i < 66) {
+        dp[i] = 0 - 1;
+        i = i + 1;
+    }
+
+    output(knapsack(n, m));
+    return 0;
+}

tests/3-codegen/autogen/testcases/13-complex.out

+15
+0

tests/3-codegen/autogen/testcases/2-calculate.cminus

+int main(void) {
+    int a;
+    int b;
+    int c;
+
+    a = 23;
+    b = 25;
+    c = 4;
+
+    return a + b * c;
+}

tests/3-codegen/autogen/testcases/2-calculate.out

+123

tests/3-codegen/autogen/testcases/3-output.cminus

+int main(void) {
+    output(11);
+    output(22222);
+    return 0;
+}

tests/3-codegen/autogen/testcases/3-output.out

+11
+22222
+0

tests/3-codegen/autogen/testcases/4-if.cminus

+int main(void) {
+    int a;
+    int b;
+    int c;
+
+    a = 11;
+    b = 22;
+    c = 33;
+
+    /* max value */
+    if (a > b) {
+        if (a > c)
+            output(a);
+        else
+            output(c);
+    } else {
+        if (c < b)
+            output(b);
+        else
+            output(c);
+    }
+    return 0;
+}

tests/3-codegen/autogen/testcases/4-if.out

+33
+0

tests/3-codegen/autogen/testcases/5-while.cminus

+int main(void) {
+    int n;
+    int i;
+
+    n = 10;
+    i = 0;
+
+    while (i < n) {
+        output(i);
+        i = i + 1;
+    }
+
+    return 0;
+}

tests/3-codegen/autogen/testcases/5-while.out

+0
+1
+2
+3
+4
+5
+6
+7
+8
+9
+0

tests/3-codegen/autogen/testcases/6-array.cminus

+int main(void) {
+    int a[10];
+    int i;
+
+    i = 0;
+    a[0] = 11;
+    a[4] = 22;
+    a[9] = 33;
+
+    output(a[0]);
+    output(a[4]);
+    output(a[9]);
+
+    return 0;
+}

tests/3-codegen/autogen/testcases/6-array.out

+11
+22
+33
+0

tests/3-codegen/autogen/testcases/7-function.cminus

+int min(int a, int b) {
+    if (a <= b)
+        return a;
+    else
+        return b;
+}
+
+int main(void) {
+    int a;
+    int b;
+    int c;
+
+    a = 11;
+    b = 22;
+    c = 33;
+
+    output(min(a, b));
+    output(min(b, c));
+    output(min(c, a));
+    return 0;
+}

tests/3-codegen/autogen/testcases/7-function.out

+11
+22
+11
+0

tests/3-codegen/autogen/testcases/8-store.cminus

+int store(int arr[], int index, int value) {
+    arr[index] = value;
+    return value;
+}
+
+int main(void) {
+    int a[10];
+    int i;
+    int sum;
+
+    i = 0;
+    while (i < 10) {
+        store(a, i, i * 2);
+        i = i + 1;
+    }
+
+    sum = 0;
+    i = 0;
+    while (i < 10) {
+        sum = sum + a[i];
+        i = i + 1;
+    }
+
+    output(sum);
+    return 0;
+}

tests/3-codegen/autogen/testcases/8-store.out

+90
+0

tests/3-codegen/autogen/testcases/9-fibonacci.cminus

+int fibonacci(int n) {
+    if (n == 0)
+        return 0;
+    else if (n == 1)
+        return 1;
+    else
+        return fibonacci(n - 1) + fibonacci(n - 2);
+}
+
+int main(void) {
+    int n;
+    int i;
+
+    n = 10;
+    i = 0;
+
+    while (i < n) {
+        output(fibonacci(i));
+        i = i + 1;
+    }
+
+    return 0;
+}

tests/3-codegen/autogen/testcases/9-fibonacci.out

+0
+1
+1
+2
+3
+5
+8
+13
+21
+34
+0

tests/3-codegen/warmup/CMakeLists.txt

+add_executable(
+    stu_assign_codegen
+    stu_cpp/assign_codegen.cpp
+)
+target_link_libraries(
+    stu_assign_codegen
+    IR_lib
+    codegen
+)
+
+add_executable(
+    stu_function_codegen
+    stu_cpp/function_codegen.cpp
+)
+target_link_libraries(
+    stu_function_codegen
+    IR_lib
+    codegen
+)
+
+add_executable(
+    stu_icmp_codegen
+    stu_cpp/icmp_codegen.cpp
+)
+target_link_libraries(
+    stu_icmp_codegen
+    IR_lib
+    codegen
+)
+
+add_executable(
+    stu_fcmp_codegen
+    stu_cpp/fcmp_codegen.cpp
+)
+target_link_libraries(
+    stu_fcmp_codegen
+    IR_lib
+    codegen
+)
+
+add_executable(
+    stu_float_codegen
+    stu_cpp/float_codegen.cpp
+)
+target_link_libraries(
+    stu_float_codegen
+    IR_lib
+    codegen
+)
+
+add_executable(
+    stu_global_codegen
+    stu_cpp/global_codegen.cpp
+)
+target_link_libraries(
+    stu_global_codegen
+    IR_lib
+    codegen
+)

tests/3-codegen/warmup/ll_cases/assign.ll

+define i32 @main() {
+label_entry:
+  %op0 = alloca [10 x i32]
+  %op1 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0
+  store i32 10, i32* %op1
+  %op2 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 1
+  %op3 = load i32, i32* %op1
+  %op4 = mul i32 %op3, 3
+  store i32 %op4, i32* %op2
+  %op5 = load i32, i32* %op2
+  ret i32 %op5
+}

tests/3-codegen/warmup/ll_cases/fcmp.ll

+define i32 @main() {
+label_entry:
+  ; 5.5 (0x40b00000) > 1.0 (0x3f800000)
+  %op0 = fcmp ugt float 0x4016000000000000, 0x3ff0000000000000
+  %op1 = zext i1 %op0 to i32
+  %op2 = icmp ne i32 %op1, 0
+  br i1 %op2, label %label3, label %label4
+label3:                                                ; preds = %label_entry
+  ret i32 233
+label4:                                                ; preds = %label_entry
+  ret i32 0
+}

tests/3-codegen/warmup/ll_cases/float.ll

+define i32 @main() {
+label_entry:
+  %op0 = alloca float
+  store float 0x40091eb860000000, float* %op0 ; 3.14, FP32 0x4048f5c3
+  %op1 = load float, float* %op0
+  %op2 = fptosi float %op1 to i32
+  ret i32 %op2
+}

tests/3-codegen/warmup/ll_cases/function.ll

+define i32 @callee(i32 %arg0) {
+label_entry:
+  %op1 = alloca i32
+  store i32 %arg0, i32* %op1
+  %op2 = load i32, i32* %op1
+  %op3 = mul i32 3, %op2
+  ret i32 %op3
+}
+
+define i32 @main() {
+label_entry:
+  %op0 = call i32 @callee(i32 110)
+  ret i32 %op0
+}

tests/3-codegen/warmup/ll_cases/global.ll

+@a = global i32 zeroinitializer
+
+define i32 @main() {
+label_entry:
+  store i32 10, i32* @a
+  %op0 = load i32, i32* @a
+  ret i32 %op0
+}

tests/3-codegen/warmup/ll_cases/icmp.ll

+define i32 @main() {
+label_entry:
+  %op0 = icmp sgt i32 5, 1
+  %op1 = zext i1 %op0 to i32
+  %op2 = icmp ne i32 %op1, 0
+  br i1 %op2, label %label3, label %label4
+label3:                                                ; preds = %label_entry
+  ret i32 233
+label4:                                                ; preds = %label_entry
+  ret i32 0
+}

tests/3-codegen/warmup/stu_cpp/assign_codegen.cpp

+#include "ASMInstruction.hpp"
+#include "CodeGen.hpp"
+#include "Module.hpp"
+
+#include <iostream>
+#include <memory>
+#include <unordered_map>
+
+void translate_main(CodeGen *codegen); // 将 main 函数翻译为汇编代码
+
+int main() {
+    auto *module = new Module();
+    auto *codegen = new CodeGen(module);
+
+    // 告诉汇编器将汇编放到代码段
+    codegen->append_inst(".text", ASMInstruction::Atrribute);
+
+    translate_main(codegen);
+
+    std::cout << codegen->print();
+    delete codegen;
+    delete module;
+    return 0;
+}
+
+// TODO: 按照提示补全
+void translate_main(CodeGen *codegen) {
+    std::unordered_map<std::string, int> offset_map;
+
+    /* 声明 main 函数 */
+    codegen->append_inst(".globl main", ASMInstruction::Atrribute);
+    codegen->append_inst(".type main, @function", ASMInstruction::Atrribute);
+
+    /* main 函数开始 */
+    codegen->append_inst("main", ASMInstruction::Label); // main 函数标签
+
+    /* main 函数的 Prologue (序言) */
+    // 保存返回地址
+    codegen->append_inst("st.d $ra, $sp, -8");
+    // 保存老的 fp
+    codegen->append_inst("st.d $fp, $sp, -16");
+    // 设置新的 fp
+    codegen->append_inst("addi.d $fp, $sp, 0");
+    // 为栈帧分配空间. 思考: 为什么是 96 字节?
+    codegen->append_inst("addi.d $sp, $sp, -96");
+
+    /* main 函数的 label_entry */
+    codegen->append_inst(".main_label_entry", ASMInstruction::Label);
+
+    /* %op0 = alloca [10 x i32] */
+    // 在汇编中写入注释, 方便 debug
+    codegen->append_inst("%op0 = alloca [10 x i32]", ASMInstruction::Comment);
+    // 将数组的地址写入 %op0 对应的内存空间中
+    // 首先计算数组的地址对应的偏移量, 按照我们在栈式分配中介绍的规则, 数组的空
+    // 间紧挨着 %op0 的空间, 数组的地址为这段空间的最小值.
+    // 然后将数组的地址写入 %op0 对应的内存空间中.
+    // 你需要思考:
+    // - %op0 的大小是多少字节?
+    // - %op0 的偏移量是多少字节?
+    // - 数组的大小是多少字节?
+    // - 数组的偏移量是多少字节?
+    offset_map["%op0"] = -24;  // TODO: 请填空
+    offset_map["array"] = -64; // TODO: 请填空
+    codegen->append_inst("addi.d",
+                         {"$t0", "$fp", std::to_string(offset_map["array"])});
+    codegen->append_inst("st.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+
+    /* %op1 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0 */
+    codegen->append_inst(
+        "%op1 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 0",
+        ASMInstruction::Comment);
+    // 将 gep 的值写入 %op1 对应的内存空间中
+    // 你可以手动计算 gep 的值, 或者使用指令计算 gep 的值
+    // 你需要思考 %op1 的偏移量是多少字节
+    offset_map["%op1"] = -72; // TODO: 请填空
+    // 获得数组的地址, 将其写入 $t0 中
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+    // TODO: 计算 %op1 的值, 并将其写入 %op1 对应的内存空间中
+    codegen->append_inst("addi.d $t0, $t0, 0");
+    codegen->append_inst("st.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+
+    /* store i32 10, i32* %op1 */
+    codegen->append_inst("store i32 10, i32* %op1", ASMInstruction::Comment);
+    // 将 10 写入数组 (%op1 指向的空间) 中, 需要先获得 %op1 的值
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+    codegen->append_inst("addi.w $t1, $zero, 10");
+    codegen->append_inst("st.w $t1, $t0, 0");
+
+    /* %op2 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 1 */
+    codegen->append_inst(
+        "%op2 = getelementptr [10 x i32], [10 x i32]* %op0, i32 0, i32 1",
+        ASMInstruction::Comment);
+    // 将 gep 的值写入 %op2 对应的内存空间中
+    // 你可以手动计算 gep 的值, 或者使用指令计算 gep 的值
+    // 你需要思考 %op2 的偏移量是多少字节
+    offset_map["%op2"] = -80; // TODO: 请填空
+    // 获得数组的地址, 将其写入 $t0 中
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+    // TODO: 计算 gep 的值, 并将其写入 %op2 对应的内存空间中
+    codegen->append_inst("addi.d $t0, $t0, 4");
+    codegen->append_inst("st.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op2"])});
+
+    /* %op3 = load i32, i32* %op1 */
+    codegen->append_inst("%op3 = load i32, i32* %op1", ASMInstruction::Comment);
+    // 将 %op1 指向的空间中存储的值写入 %op3 对应的内存空间中
+    offset_map["%op3"] = -84; // TODO: 请填空
+    // TODO: 先获得 %op1 的值, 然后获得 %op1 指向的空间的值, 最后将这个值写入
+    // %op3 对应的内存空间中
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+    codegen->append_inst("ld.w $t1, $t0, 0");
+    codegen->append_inst("st.w",
+                         {"$t1", "$fp", std::to_string(offset_map["%op3"])});
+
+    /* %op4 = mul i32 %op3, 3 */
+    codegen->append_inst("%op4 = mul i32 %op3, 3", ASMInstruction::Comment);
+    offset_map["%op4"] = -88; // TODO: 请填空
+    // TODO: 先获得 %op3 的值, 然后将其乘以 3, 最后将结果写入 %op4 对应的内存空
+    // 间中
+    codegen->append_inst("ld.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op3"])});
+    codegen->append_inst("addi.w $t1, $zero, 3");
+    codegen->append_inst("mul.w $t0, $t0, $t1");
+    codegen->append_inst("st.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op4"])});
+
+    /* store i32 %op4, i32* %op2 */
+    codegen->append_inst("store i32 %op4, i32* %op2", ASMInstruction::Comment);
+    // 将 %op4 的值写入 %op2 指向的空间中
+    // TODO: 先获得 %op4 的值, 然后将其写入 %op2 指向的空间中
+    codegen->append_inst("ld.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op4"])});
+    codegen->append_inst("ld.d",
+                         {"$t1", "$fp", std::to_string(offset_map["%op2"])});
+    codegen->append_inst("st.w $t0, $t1, 0");
+
+    /* %op5 = load i32, i32* %op2 */
+    // 将 %op2 指向的空间中存储的值写入 %op5 对应的内存空间中
+    offset_map["%op5"] = -92; // TODO: 请填空
+    // TODO: 先获得 %op2 的值, 然后获得 %op2 指向的空间的值, 最后将这个值写入
+    // %op5 对应的内存空间中
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op2"])});
+    codegen->append_inst("ld.w $t1, $t0, 0");
+    codegen->append_inst("st.w",
+                         {"$t1", "$fp", std::to_string(offset_map["%op5"])});
+
+    /* ret i32 %op5 */
+    codegen->append_inst("ret i32 %op5", ASMInstruction::Comment);
+    // TODO: 将 %op5 的值写入 $a0 中
+    codegen->append_inst("ld.w",
+                         {"$a0", "$fp", std::to_string(offset_map["%op5"])});
+    // 思考: 为什么不在这里 jr $ra, 而是跳转到 main_exit?
+    codegen->append_inst("b main_exit");
+
+    /* main 函数的 Epilogue (收尾) */
+    codegen->append_inst("main_exit", ASMInstruction::Label);
+    // 释放栈帧空间
+    codegen->append_inst("addi.d $sp, $sp, 96");
+    // 恢复 ra
+    codegen->append_inst("ld.d $ra, $sp, -8");
+    // 恢复 fp
+    codegen->append_inst("ld.d $fp, $sp, -16");
+    // 返回
+    codegen->append_inst("jr $ra");
+}

tests/3-codegen/warmup/stu_cpp/fcmp_codegen.cpp

+#include "ASMInstruction.hpp"
+#include "CodeGen.hpp"
+#include "Module.hpp"
+
+#include <iostream>
+#include <memory>
+#include <unordered_map>
+
+void translate_main(CodeGen *codegen); // 将 main 函数翻译为汇编代码
+
+int main() {
+    auto *module = new Module();
+    auto *codegen = new CodeGen(module);
+
+    // 告诉汇编器将汇编放到代码段
+    codegen->append_inst(".text", ASMInstruction::Atrribute);
+
+    translate_main(codegen);
+
+    std::cout << codegen->print();
+    delete codegen;
+    delete module;
+    return 0;
+}
+
+// TODO: 按照提示补全
+void translate_main(CodeGen *codegen) {
+    std::unordered_map<std::string, int> offset_map;
+
+    /* 声明 main 函数 */
+    codegen->append_inst(".globl main", ASMInstruction::Atrribute);
+    codegen->append_inst(".type main, @function", ASMInstruction::Atrribute);
+
+    /* main 函数开始 */
+    codegen->append_inst("main", ASMInstruction::Label); // main 函数标签
+
+    /* main 函数的 Prologue (序言) */
+    // 保存返回地址
+    codegen->append_inst("st.d $ra, $sp, -8");
+    // 保存老的 fp
+    codegen->append_inst("st.d $fp, $sp, -16");
+    // 设置新的 fp
+    codegen->append_inst("addi.d $fp, $sp, 0");
+    // 为栈帧分配空间. 思考: 为什么是 32 字节?
+    codegen->append_inst("addi.d $sp, $sp, -32");
+
+    /* main 函数的 label_entry */
+    codegen->append_inst(".main_label_entry", ASMInstruction::Label);
+
+    /* %op0 = fcmp ugt float 0x4016000000000000, 0x3ff0000000000000 */
+    // 在汇编中写入注释, 方便 debug
+    codegen->append_inst(
+        "%op0 = fcmp ugt float 0x4016000000000000, 0x3ff0000000000000",
+        ASMInstruction::Comment);
+    // 将比较结果写入 %op0 对应的内存空间中
+    offset_map["%op0"] = -17; // TODO: 请填空
+    // TODO: 将 5.5 (0x40b00000) 加载到浮点寄存器中
+    codegen->append_inst("lu12i.w $t0, 0x40b00");
+    codegen->append_inst("movgr2fr.w $ft0, $t0");
+    // TODO: 将 1.0 (0x3f800000) 加载到浮点寄存器中
+    codegen->append_inst("lu12i.w $t0, 0x3f80");
+    codegen->append_inst("movgr2fr.w $ft1, $t0");
+    // TODO: 使用 fcmp.slt.s 进行比较, 比较结果在浮点标志寄存器中, 你需要思考
+    // 如何将浮点标志寄存器中的值写入内存. 提示: 尝试使用 bcnez 指令
+    codegen->append_inst("fcmp.slt.s $fcc0, $ft1, $ft0");
+    codegen->append_inst("addi.w $t0, $zero, 1");
+    codegen->append_inst("bcnez $fcc0, .main_label_entry_fcmp_0");
+    codegen->append_inst("addi.w $t0, $zero, 0");
+    codegen->append_inst(".main_label_entry_fcmp_0", ASMInstruction::Label);
+    codegen->append_inst("st.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+
+    /* %op1 = zext i1 %op0 to i32 */
+    codegen->append_inst("%op1 = zext i1 %op0 to i32", ASMInstruction::Comment);
+    // 将 %op0 的值从 i1 类型转换为 i32 类型, 并将结果写入到 %op1 对应的内存空
+    // 间中
+    offset_map["%op1"] = -21; // TODO: 请填空
+    // TODO: 获得 %op0 的值, 然后进行转换, 最后将结果写入 %op1
+    // 思考: 怎么转换? 需不需要显式地使用某些指令转换?
+    codegen->append_inst("ld.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+    codegen->append_inst("st.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+
+    /* %op2 = icmp ne i32 %op1, 0 */
+    codegen->append_inst("%op2 = icmp ne i32 %op1, 0", ASMInstruction::Comment);
+    // 比较 %op1 和 0, 并将结果写入 %op2 对应的内存空间中
+    offset_map["%op2"] = -22; // TODO: 请填空
+    // TODO: 获得 %op1 的值, 然后进行比较, 最后将结果写入 %op2
+    // 思考: 如何比较? 能否不使用跳转指令计算结果?
+    // 提示: 尝试使用 xor/xori 和 slt/sltu/slti/sltui 计算比较结果
+    codegen->append_inst("ld.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+    codegen->append_inst("xori $t0, $t0, 0");
+    codegen->append_inst("sltu $t0, $zero, $t0");
+    codegen->append_inst("st.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op2"])});
+
+    /* br i1 %op2, label %label3, label %label4 */
+    codegen->append_inst("br i1 %op2, label %label3, label %label4",
+                         ASMInstruction::Comment);
+    // TODO: 获得 %op2 的值, 并根据 %op2 的值跳转到 label3 或者 label4
+    // 提示: 汇编中对应的标签分别为 .main_label3 和 .main_label4
+    codegen->append_inst("ld.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op2"])});
+    codegen->append_inst("bne $t0, $zero, .main_label3");
+    codegen->append_inst("b .main_label4");
+
+    /* label3: */
+    codegen->append_inst(".main_label3", ASMInstruction::Label);
+
+    /* ret i32 233 */
+    codegen->append_inst("ret i32 233", ASMInstruction::Comment);
+    // 将 233 写入 $a0 中
+    codegen->append_inst("addi.w $a0, $zero, 233");
+    codegen->append_inst("b main_exit");
+
+    /* label4: */
+    codegen->append_inst(".main_label4", ASMInstruction::Label);
+
+    /* ret i32 0 */
+    codegen->append_inst("ret i32 0", ASMInstruction::Comment);
+    // 将 0 写入 $a0 中
+    codegen->append_inst("addi.w $a0, $zero, 0");
+    codegen->append_inst("b main_exit");
+
+    /* main 函数的 Epilogue (收尾) */
+    codegen->append_inst("main_exit", ASMInstruction::Label);
+    // 释放栈帧空间
+    codegen->append_inst("addi.d $sp, $sp, 32");
+    // 恢复 ra
+    codegen->append_inst("ld.d $ra, $sp, -8");
+    // 恢复 fp
+    codegen->append_inst("ld.d $fp, $sp, -16");
+    // 返回
+    codegen->append_inst("jr $ra");
+}

tests/3-codegen/warmup/stu_cpp/float_codegen.cpp

+#include "ASMInstruction.hpp"
+#include "CodeGen.hpp"
+#include "Module.hpp"
+
+#include <iostream>
+#include <memory>
+#include <unordered_map>
+
+void translate_main(CodeGen *codegen); // 将 main 函数翻译为汇编代码
+
+int main() {
+    auto *module = new Module();
+    auto *codegen = new CodeGen(module);
+
+    // 告诉汇编器将汇编放到代码段
+    codegen->append_inst(".text", ASMInstruction::Atrribute);
+
+    translate_main(codegen);
+
+    std::cout << codegen->print();
+    delete codegen;
+    delete module;
+    return 0;
+}
+
+// TODO: 按照提示补全
+void translate_main(CodeGen *codegen) {
+    std::unordered_map<std::string, int> offset_map;
+
+    /* 声明 main 函数 */
+    codegen->append_inst(".globl main", ASMInstruction::Atrribute);
+    codegen->append_inst(".type main, @function", ASMInstruction::Atrribute);
+
+    /* main 函数开始 */
+    codegen->append_inst("main", ASMInstruction::Label); // main 函数标签
+
+    /* main 函数的 Prologue (序言) */
+    // 保存返回地址
+    codegen->append_inst("st.d $ra, $sp, -8");
+    // 保存老的 fp
+    codegen->append_inst("st.d $fp, $sp, -16");
+    // 设置新的 fp
+    codegen->append_inst("addi.d $fp, $sp, 0");
+    // 为栈帧分配空间. 思考: 为什么是 48 字节?
+    codegen->append_inst("addi.d $sp, $sp, -48");
+
+    /* main 函数的 label_entry */
+    codegen->append_inst(".main_label_entry", ASMInstruction::Label);
+
+    /* %op0 = alloca float */
+    // 在汇编中写入注释, 方便 debug
+    codegen->append_inst("%op0 = alloca float", ASMInstruction::Comment);
+    // 将浮点数的地址写入 %op0 对应的内存空间中
+    offset_map["%op0"] = -24;  // TODO: 请填空
+    offset_map["*%op0"] = -28; // TODO: 请填空
+    codegen->append_inst("addi.d",
+                         {"$t0", "$fp", std::to_string(offset_map["*%op0"])});
+    codegen->append_inst("st.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+
+    /* store float 0x40091eb860000000, float* %op0 */
+    codegen->append_inst("store float 0x40091eb860000000, float* %op0",
+                         ASMInstruction::Comment);
+    // 将 3.14 (0x4048f5c3) 写入 %op0 指向的内存空间中
+    // 获得 %op0 的值
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+    // TODO: 将 0x4048f5c3 加载到通用寄存器或者浮点寄存器中
+    codegen->append_inst("lu12i.w $t1, 0x4048f");
+    codegen->append_inst("ori $t1, $t1, 0x5c3");
+    // TODO: 将通用寄存器或者浮点寄存器中的值写入 %op0 对应的内存空间中
+    codegen->append_inst("st.w $t1, $t0, 0");
+
+    /* %op1 = load float, float* %op0 */
+    codegen->append_inst("%op1 = load float, float* %op0",
+                         ASMInstruction::Comment);
+    // TODO: 先获得 %op0 的值, 然后获得 %op0 指向的空间的值, 最后将这个值写入
+    // %op1 对应的内存空间中
+    offset_map["%op1"] = -32; // TODO: 请填空
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+    codegen->append_inst("fld.s $ft0, $t0, 0");
+    codegen->append_inst("fst.s",
+                         {"$ft0", "$fp", std::to_string(offset_map["%op1"])});
+
+    /* %op2 = fptosi float %op1 to i32 */
+    codegen->append_inst("%op2 = fptosi float %op1 to i32",
+                         ASMInstruction::Comment);
+    // TODO: 使用 ftintrz.w.s 指令进行转换, 并将结果写入 %op2 对应的内存空间中
+    offset_map["%op2"] = -36; // TODO: 请填空
+    codegen->append_inst("fld.s",
+                         {"$ft0", "$fp", std::to_string(offset_map["%op1"])});
+    codegen->append_inst("ftintrz.w.s $ft0, $ft0");
+    codegen->append_inst("fst.s",
+                         {"$ft0", "$fp", std::to_string(offset_map["%op2"])});
+
+    /* ret i32 %op2 */
+    codegen->append_inst("ret i32 %op2", ASMInstruction::Comment);
+    // TODO: 将 %op2 的值写入 $a0 中
+    codegen->append_inst("ld.w",
+                         {"$a0", "$fp", std::to_string(offset_map["%op2"])});
+    codegen->append_inst("b main_exit");
+
+    /* main 函数的 Epilogue (收尾) */
+    codegen->append_inst("main_exit", ASMInstruction::Label);
+    // 释放栈帧空间
+    codegen->append_inst("addi.d $sp, $sp, 48");
+    // 恢复 ra
+    codegen->append_inst("ld.d $ra, $sp, -8");
+    // 恢复 fp
+    codegen->append_inst("ld.d $fp, $sp, -16");
+    // 返回
+    codegen->append_inst("jr $ra");
+}

tests/3-codegen/warmup/stu_cpp/function_codegen.cpp

+#include "ASMInstruction.hpp"
+#include "CodeGen.hpp"
+#include "Module.hpp"
+
+#include <iostream>
+#include <memory>
+#include <unordered_map>
+
+void translate_main(CodeGen *codegen);   // 将 main 函数翻译为汇编代码
+void translate_callee(CodeGen *codegen); // 将 callee 函数翻译为汇编代码
+
+int main() {
+    auto *module = new Module();
+    auto *codegen = new CodeGen(module);
+
+    // 告诉汇编器将汇编放到代码段
+    codegen->append_inst(".text", ASMInstruction::Atrribute);
+
+    translate_callee(codegen);
+    translate_main(codegen);
+
+    std::cout << codegen->print();
+    delete codegen;
+    delete module;
+    return 0;
+}
+
+// TODO: 按照提示补全
+void translate_callee(CodeGen *codegen) {
+    std::unordered_map<std::string, int> offset_map;
+
+    /* 声明 callee 函数 */
+    codegen->append_inst(".globl callee", ASMInstruction::Atrribute);
+    codegen->append_inst(".type callee, @function", ASMInstruction::Atrribute);
+
+    /* callee 函数开始 */
+    codegen->append_inst("callee", ASMInstruction::Label); // main 函数标签
+
+    /* callee 函数的 Prologue (序言) */
+    // 保存返回地址
+    codegen->append_inst("st.d $ra, $sp, -8");
+    // 保存老的 fp
+    codegen->append_inst("st.d $fp, $sp, -16");
+    // 设置新的 fp
+    codegen->append_inst("addi.d $fp, $sp, 0");
+    // 为栈帧分配空间. 思考: 为什么是 48 字节?
+    codegen->append_inst("addi.d $sp, $sp, -48");
+
+    // 为参数分配空间
+    offset_map["%arg0"] = -20; // TODO: 请填空
+    codegen->append_inst("st.w",
+                         {"$a0", "$fp", std::to_string(offset_map["%arg0"])});
+
+    /* callee 函数的 label_entry */
+    codegen->append_inst(".callee_label_entry", ASMInstruction::Label);
+
+    /* %op1 = alloca i32 */
+    // 在汇编中写入注释, 方便 debug
+    codegen->append_inst("%op1 = alloca i32", ASMInstruction::Comment);
+    // 将 alloca 的地址写入 %op1 对应的内存空间中
+    offset_map["%op1"] = -28;  // TODO: 请填空
+    offset_map["*%op1"] = -32; // TODO: 请填空
+    codegen->append_inst("addi.d",
+                         {"$t0", "$fp", std::to_string(offset_map["*%op1"])});
+    codegen->append_inst("st.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+
+    /* store i32 %arg0, i32* %op1 */
+    codegen->append_inst("store i32 %arg0, i32* %op1", ASMInstruction::Comment);
+    // 将 %arg0 的值写入 %op1 对应的内存空间中
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+    codegen->append_inst("ld.w",
+                         {"$t1", "$fp", std::to_string(offset_map["%arg0"])});
+    codegen->append_inst("st.w $t1, $t0, 0");
+
+    /* %op2 = load i32, i32* %op1 */
+    codegen->append_inst("%op2 = load i32, i32* %op1", ASMInstruction::Comment);
+    // 将 %op1 对应的内存空间的值写入 %op2 对应的内存空间中
+    offset_map["%op2"] = -36; // TODO: 请填空
+    codegen->append_inst("ld.d",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+    codegen->append_inst("ld.w $t1, $t0, 0");
+    codegen->append_inst("st.w",
+                         {"$t1", "$fp", std::to_string(offset_map["%op2"])});
+
+    /* %op3 = mul i32 3, %op2 */
+    codegen->append_inst("%op3 = mul i32 3, %op2", ASMInstruction::Comment);
+    // 将 %op2 的值乘以 3, 并将结果写入 %op3 对应的内存空间中
+    offset_map["%op3"] = -40; // TODO: 请填空
+    codegen->append_inst("ld.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op2"])});
+    codegen->append_inst("addi.w $t1, $zero, 3");
+    codegen->append_inst("mul.w $t0, $t0, $t1");
+    codegen->append_inst("st.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op3"])});
+
+    /* ret i32 %op3 */
+    codegen->append_inst("ret i32 %op3", ASMInstruction::Comment);
+    // 将 %op3 的值写入 $a0 中
+    codegen->append_inst("ld.w",
+                         {"$a0", "$fp", std::to_string(offset_map["%op3"])});
+    codegen->append_inst("b callee_exit");
+
+    /* callee 函数的 Epilogue (收尾) */
+    codegen->append_inst("callee_exit", ASMInstruction::Label);
+    // 释放栈帧空间
+    codegen->append_inst("addi.d $sp, $sp, 48");
+    // 恢复 ra
+    codegen->append_inst("ld.d $ra, $sp, -8");
+    // 恢复 fp
+    codegen->append_inst("ld.d $fp, $sp, -16");
+    // 返回
+    codegen->append_inst("jr $ra");
+}
+
+// TODO: 按照提示补全
+void translate_main(CodeGen *codegen) {
+    std::unordered_map<std::string, int> offset_map;
+
+    /* 声明 main 函数 */
+    codegen->append_inst(".globl main", ASMInstruction::Atrribute);
+    codegen->append_inst(".type main, @function", ASMInstruction::Atrribute);
+
+    /* main 函数开始 */
+    codegen->append_inst("main", ASMInstruction::Label); // main 函数标签
+
+    /* main 函数的 Prologue (序言) */
+    // 保存返回地址
+    codegen->append_inst("st.d $ra, $sp, -8");
+    // 保存老的 fp
+    codegen->append_inst("st.d $fp, $sp, -16");
+    // 设置新的 fp
+    codegen->append_inst("addi.d $fp, $sp, 0");
+    // 为栈帧分配空间. 思考: 为什么是 32 字节?
+    codegen->append_inst("addi.d $sp, $sp, -32");
+
+    /* main 函数的 label_entry */
+    codegen->append_inst(".main_label_entry", ASMInstruction::Label);
+
+    /* %op0 = call i32 @callee(i32 110) */
+    // 在汇编中写入注释, 方便 debug
+    codegen->append_inst("%op0 = call i32 @callee(i32 110)",
+                         ASMInstruction::Comment);
+    offset_map["%op0"] = -20; // TODO: 请填空
+    // TODO: 将参数写入参数寄存器
+    codegen->append_inst("addi.w $a0, $zero, 110");
+    // TODO: 调用 callee 函数
+    codegen->append_inst("bl callee");
+    // 将返回值写入 %op0 对应的内存空间中
+    codegen->append_inst("st.w",
+                         {"$a0", "$fp", std::to_string(offset_map["%op0"])});
+
+    /* ret i32 %op0 */
+    codegen->append_inst("ret i32 %op0", ASMInstruction::Comment);
+    codegen->append_inst("ld.w",
+                         {"$a0", "$fp", std::to_string(offset_map["%op0"])});
+    codegen->append_inst("b main_exit");
+
+    /* main 函数的 Epilogue (收尾) */
+    codegen->append_inst("main_exit", ASMInstruction::Label);
+    // 释放栈帧空间
+    codegen->append_inst("addi.d $sp, $sp, 32");
+    // 恢复 ra
+    codegen->append_inst("ld.d $ra, $sp, -8");
+    // 恢复 fp
+    codegen->append_inst("ld.d $fp, $sp, -16");
+    // 返回
+    codegen->append_inst("jr $ra");
+}

tests/3-codegen/warmup/stu_cpp/global_codegen.cpp

+#include "ASMInstruction.hpp"
+#include "CodeGen.hpp"
+#include "Module.hpp"
+
+#include <iostream>
+#include <memory>
+#include <unordered_map>
+
+void declare_global(CodeGen *codegen); // 声明全局变量 a
+void translate_main(CodeGen *codegen); // 将 main 函数翻译为汇编代码
+
+int main() {
+    auto *module = new Module();
+    auto *codegen = new CodeGen(module);
+
+    // 告诉汇编器将汇编放到代码段
+    codegen->append_inst(".text", ASMInstruction::Atrribute);
+
+    declare_global(codegen);
+
+    codegen->append_inst(".text", ASMInstruction::Atrribute);
+    translate_main(codegen);
+
+    std::cout << codegen->print();
+    delete codegen;
+    delete module;
+    return 0;
+}
+
+void declare_global(CodeGen *codegen) {
+    // 声明全局变量 a
+    // 将 a 放到 .bss 段中
+    codegen->append_inst(".section .bss, \"aw\", @nobits",
+                         ASMInstruction::Atrribute);
+    // 标记 a 为全局变量
+    codegen->append_inst(".globl a", ASMInstruction::Atrribute);
+    // 标记 a 为数据对象 (变量)
+    codegen->append_inst(".type a, @object", ASMInstruction::Atrribute);
+    // 标记 a 的大小为 4 字节
+    codegen->append_inst(".size a, 4", ASMInstruction::Atrribute);
+    // a 的标签
+    codegen->append_inst("a", ASMInstruction::Label);
+    // 为 a 分配空间
+    codegen->append_inst(".space 4");
+}
+
+// TODO: 按照提示补全
+void translate_main(CodeGen *codegen) {
+    std::unordered_map<std::string, int> offset_map;
+
+    /* 声明 main 函数 */
+    codegen->append_inst(".globl main", ASMInstruction::Atrribute);
+    codegen->append_inst(".type main, @function", ASMInstruction::Atrribute);
+
+    /* main 函数开始 */
+    codegen->append_inst("main", ASMInstruction::Label); // main 函数标签
+
+    /* main 函数的 Prologue (序言) */
+    // 保存返回地址
+    codegen->append_inst("st.d $ra, $sp, -8");
+    // 保存老的 fp
+    codegen->append_inst("st.d $fp, $sp, -16");
+    // 设置新的 fp
+    codegen->append_inst("addi.d $fp, $sp, 0");
+    // 为栈帧分配空间. 思考: 为什么是 32 字节?
+    codegen->append_inst("addi.d $sp, $sp, -32");
+
+    /* main 函数的 label_entry */
+    codegen->append_inst(".main_label_entry", ASMInstruction::Label);
+
+    /* store i32 10, i32* @a */
+    // 在汇编中写入注释, 方便 debug
+    codegen->append_inst("store i32 10, i32* @a", ASMInstruction::Comment);
+    // 将 10 写入 a 对应的内存空间中
+    // TODO: 获得 a 的地址
+    codegen->append_inst("la.local $t0, a");
+    // TODO: 将 10 写入 a 对应的内存空间中
+    codegen->append_inst("addi.w $t1, $zero, 10");
+    codegen->append_inst("st.w $t1, $t0, 0");
+
+    /* %op0 = load i32, i32* @a */
+    codegen->append_inst("%op0 = load i32, i32* @a", ASMInstruction::Comment);
+    // 将 a 的值写入 %op0 对应的内存空间中
+    offset_map["%op0"] = -20; // TODO: 请填空
+    // TODO: 获得 a 的地址, 并存储在 $t0 中
+    codegen->append_inst("la.local $t0, a");
+    // 将 a 的值写入 %op0 对应的内存空间中
+    codegen->append_inst("ld.w $t1, $t0, 0");
+    codegen->append_inst("st.w",
+                         {"$t1", "$fp", std::to_string(offset_map["%op0"])});
+
+    /* ret i32 %op0 */
+    codegen->append_inst("ret i32 %op0", ASMInstruction::Comment);
+    // 将 %op0 的值写入 $a0
+    codegen->append_inst("ld.w",
+                         {"$a0", "$fp", std::to_string(offset_map["%op0"])});
+    codegen->append_inst("b main_exit");
+
+    /* main 函数的 Epilogue (收尾) */
+    codegen->append_inst("main_exit", ASMInstruction::Label);
+    // 释放栈帧空间
+    codegen->append_inst("addi.d $sp, $sp, 32");
+    // 恢复 ra
+    codegen->append_inst("ld.d $ra, $sp, -8");
+    // 恢复 fp
+    codegen->append_inst("ld.d $fp, $sp, -16");
+    // 返回
+    codegen->append_inst("jr $ra");
+}

tests/3-codegen/warmup/stu_cpp/icmp_codegen.cpp

+#include "ASMInstruction.hpp"
+#include "CodeGen.hpp"
+#include "Module.hpp"
+
+#include <iostream>
+#include <memory>
+#include <unordered_map>
+
+void translate_main(CodeGen *codegen); // 将 main 函数翻译为汇编代码
+
+int main() {
+    auto *module = new Module();
+    auto *codegen = new CodeGen(module);
+
+    // 告诉汇编器将汇编放到代码段
+    codegen->append_inst(".text", ASMInstruction::Atrribute);
+
+    translate_main(codegen);
+
+    std::cout << codegen->print();
+    delete codegen;
+    delete module;
+    return 0;
+}
+
+// TODO: 按照提示补全
+void translate_main(CodeGen *codegen) {
+    std::unordered_map<std::string, int> offset_map;
+
+    /* 声明 main 函数 */
+    codegen->append_inst(".globl main", ASMInstruction::Atrribute);
+    codegen->append_inst(".type main, @function", ASMInstruction::Atrribute);
+
+    /* main 函数开始 */
+    codegen->append_inst("main", ASMInstruction::Label); // main 函数标签
+
+    /* main 函数的 Prologue (序言) */
+    // 保存返回地址
+    codegen->append_inst("st.d $ra, $sp, -8");
+    // 保存老的 fp
+    codegen->append_inst("st.d $fp, $sp, -16");
+    // 设置新的 fp
+    codegen->append_inst("addi.d $fp, $sp, 0");
+    // 为栈帧分配空间. 思考: 为什么是 32 字节?
+    codegen->append_inst("addi.d $sp, $sp, -32");
+
+    /* main 函数的 label_entry */
+    codegen->append_inst(".main_label_entry", ASMInstruction::Label);
+
+    /* %op0 = icmp sgt i32 5, 1 */
+    // 在汇编中写入注释, 方便 debug
+    codegen->append_inst("%op0 = icmp sgt i32 5, 1", ASMInstruction::Comment);
+    // 将比较结果写入 %op0 对应的内存空间中
+    offset_map["%op0"] = -17; // TODO: 请填空
+    codegen->append_inst("addi.w $t0, $zero, 5");
+    codegen->append_inst("addi.w $t1, $zero, 1");
+    // TODO: 使用 slt 指令比较
+    codegen->append_inst("slt $t0, $t1, $t0");
+    codegen->append_inst("st.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+
+    /* %op1 = zext i1 %op0 to i32 */
+    codegen->append_inst("%op1 = zext i1 %op0 to i32", ASMInstruction::Comment);
+    // 将 %op0 的值从 i1 类型转换为 i32 类型, 并将结果写入到 %op1 对应的内存空
+    // 间中
+    offset_map["%op1"] = -21; // TODO: 请填空
+    // TODO: 获得 %op0 的值, 然后进行转换, 最后将结果写入 %op1
+    // 思考: 怎么转换? 需不需要显式地使用某些指令转换?
+    codegen->append_inst("ld.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op0"])});
+    codegen->append_inst("st.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+
+    /* %op2 = icmp ne i32 %op1, 0 */
+    codegen->append_inst("%op2 = icmp ne i32 %op1, 0", ASMInstruction::Comment);
+    // 比较 %op1 和 0, 并将结果写入 %op2 对应的内存空间中
+    offset_map["%op2"] = -22; // TODO: 请填空
+    // TODO: 获得 %op1 的值, 然后进行比较, 最后将结果写入 %op2
+    // 思考: 如何比较? 能否不使用跳转指令计算结果?
+    // 提示: 尝试使用 xor/xori 和 slt/sltu/slti/sltui 计算比较结果
+    codegen->append_inst("ld.w",
+                         {"$t0", "$fp", std::to_string(offset_map["%op1"])});
+    codegen->append_inst("xori $t0, $t0, 0");
+    codegen->append_inst("sltu $t0, $zero, $t0");
+    codegen->append_inst("st.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op2"])});
+
+    /* br i1 %op2, label %label3, label %label4 */
+    codegen->append_inst("br i1 %op2, label %label3, label %label4",
+                         ASMInstruction::Comment);
+    // TODO: 获得 %op2 的值, 并根据 %op2 的值跳转到 label3 或者 label4
+    // 提示: 汇编中对应的标签分别为 .main_label3 和 .main_label4
+    codegen->append_inst("ld.b",
+                         {"$t0", "$fp", std::to_string(offset_map["%op2"])});
+    codegen->append_inst("bne $t0, $zero, .main_label3");
+    codegen->append_inst("b .main_label4");
+
+    /* label3: */
+    codegen->append_inst(".main_label3", ASMInstruction::Label);
+
+    /* ret i32 233 */
+    codegen->append_inst("ret i32 233", ASMInstruction::Comment);
+    // 将 233 写入 $a0 中
+    codegen->append_inst("addi.w $a0, $zero, 233");
+    codegen->append_inst("b main_exit");
+
+    /* label4: */
+    codegen->append_inst(".main_label4", ASMInstruction::Label);
+
+    /* ret i32 0 */
+    codegen->append_inst("ret i32 0", ASMInstruction::Comment);
+    // 将 0 写入 $a0 中
+    codegen->append_inst("addi.w $a0, $zero, 0");
+    codegen->append_inst("b main_exit");
+
+    /* main 函数的 Epilogue (收尾) */
+    codegen->append_inst("main_exit", ASMInstruction::Label);
+    // 释放栈帧空间
+    codegen->append_inst("addi.d $sp, $sp, 32");
+    // 恢复 ra
+    codegen->append_inst("ld.d $ra, $sp, -8");
+    // 恢复 fp
+    codegen->append_inst("ld.d $fp, $sp, -16");
+    // 返回
+    codegen->append_inst("jr $ra");
+}

tests/CMakeLists.txt

 add_subdirectory("2-ir-gen/warmup")
+add_subdirectory("3-codegen/warmup")
\ No newline at end of file