// SPDX-License-Identifier: GPL-2.0+ /******************************************************************************* * QtMips - MIPS 32-bit Architecture Subset Simulator * * Implemented to support following courses: * * B35APO - Computer Architectures * https://cw.fel.cvut.cz/wiki/courses/b35apo * * B4M35PAP - Advanced Computer Architectures * https://cw.fel.cvut.cz/wiki/courses/b4m35pap/start * * Copyright (c) 2017-2019 Karel Koci * Copyright (c) 2019 Pavel Pisa * * Faculty of Electrical Engineering (http://www.fel.cvut.cz) * Czech Technical University (http://www.cvut.cz/) * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version 2 * of the License, or (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, * Boston, MA 02110-1301, USA. * ******************************************************************************/ #include #include "tst_machine.h" #include "core.h" #include "cache.h" #include "machineconfig.h" using namespace machine; static void core_regs_data() { QTest::addColumn("i"); QTest::addColumn("init"); QTest::addColumn("res"); // Note that we shouldn't be touching program counter as that is handled automatically and differs if we use pipelining // Arithmetic instructions { Registers regs_init; regs_init.write_gp(24, 24); regs_init.write_gp(25, 12); Registers regs_res(regs_init); regs_res.write_gp(26, 36); QTest::newRow("ADD") << Instruction(0, 24, 25, 26, 0, 32) \ << regs_init \ << regs_res; QTest::newRow("ADDU") << Instruction(0, 24, 25, 26, 0, 33) \ << regs_init \ << regs_res; QTest::newRow("ADDI") << Instruction(8, 24, 26, 12) \ << regs_init \ << regs_res; QTest::newRow("ADDIU") << Instruction(9, 24, 26, 12) \ << regs_init \ << regs_res; regs_res.write_gp(26, 12); QTest::newRow("SUB") << Instruction(0, 24, 25, 26, 0, 34) \ << regs_init \ << regs_res; QTest::newRow("SUBU") << Instruction(0, 24, 25, 26, 0, 35) \ << regs_init \ << regs_res; } { Registers regs_init; regs_init.write_gp(24, 12); regs_init.write_gp(25, 24); Registers regs_res(regs_init); regs_res.write_gp(26, 1); QTest::newRow("SLT") << Instruction(0, 24, 25, 26, 0, 42) \ << regs_init \ << regs_res; QTest::newRow("SLTU") << Instruction(0, 24, 25, 26, 0, 43) \ << regs_init \ << regs_res; QTest::newRow("SLTI") << Instruction(10, 24, 26, 24) \ << regs_init \ << regs_res; QTest::newRow("SLTIU") << Instruction(11, 24, 26, 24) \ << regs_init \ << regs_res; } // Shift instructions { Registers regs_init; regs_init.write_gp(24, 0xf0); regs_init.write_gp(25, 3); Registers regs_res(regs_init); regs_res.write_gp(26, 0x780); QTest::newRow("SLL") << Instruction(0, 0, 24, 26, 3, 0) \ << regs_init \ << regs_res; QTest::newRow("SLLV") << Instruction(0, 25, 24, 26, 0, 4) \ << regs_init \ << regs_res; regs_res.write_gp(26, 0x1e); QTest::newRow("SLR") << Instruction(0, 0, 24, 26, 3, 2) \ << regs_init \ << regs_res; QTest::newRow("SLRV") << Instruction(0, 25, 24, 26, 0, 6) \ << regs_init \ << regs_res; } { Registers regs_init; regs_init.write_gp(24, 0x800000f0); regs_init.write_gp(25, 3); Registers regs_res(regs_init); regs_res.write_gp(26, 0xF000001e); QTest::newRow("SRA") << Instruction(0, 0, 24, 26, 3, 3) \ << regs_init \ << regs_res; QTest::newRow("SRAV") << Instruction(0, 25, 24, 26, 0, 7) \ << regs_init \ << regs_res; } // Logical instructions { Registers regs_init; regs_init.write_gp(24, 0xf0); regs_init.write_gp(25, 0xe1); Registers regs_res(regs_init); regs_res.write_gp(26, 0xe0); QTest::newRow("AND") << Instruction(0, 24, 25, 26, 0, 36) \ << regs_init \ << regs_res; QTest::newRow("ANDI") << Instruction(12, 24, 26, 0xe1) \ << regs_init \ << regs_res; regs_res.write_gp(26, 0xf1); QTest::newRow("OR") << Instruction(0, 24, 25, 26, 0, 37) \ << regs_init \ << regs_res; QTest::newRow("ORI") << Instruction(13, 24, 26, 0xe1) \ << regs_init \ << regs_res; regs_res.write_gp(26, 0x11); QTest::newRow("XOR") << Instruction(0, 24, 25, 26, 0, 38) \ << regs_init \ << regs_res; QTest::newRow("XORI") << Instruction(14, 24, 26, 0xe1) \ << regs_init \ << regs_res; regs_res.write_gp(26, 0xffffff0e); QTest::newRow("NOR") << Instruction(0, 24, 25, 26, 0, 39) \ << regs_init \ << regs_res; regs_res.write_gp(26, 0xf00f0000); QTest::newRow("LUI") << Instruction(15, 0, 26, 0xf00f) \ << regs_init \ << regs_res; } // Move instructions { Registers regs_init; regs_init.write_hi_lo(true, 24); regs_init.write_hi_lo(false, 28); regs_init.write_gp(24, 55); regs_init.write_gp(25, 56); regs_init.write_gp(27, 21); regs_init.write_gp(28, 22); Registers regs_res(regs_init); regs_res.write_gp(26, 24); QTest::newRow("MFHI") << Instruction(0, 0, 0, 26, 0, 16) \ << regs_init \ << regs_res; regs_res.write_gp(26, 28); QTest::newRow("MFLO") << Instruction(0, 0, 0, 26, 0, 18) \ << regs_init \ << regs_res; regs_res.write_gp(26, 0); regs_res.write_hi_lo(true, 21); QTest::newRow("MTHI") << Instruction(0, 27, 0, 0, 0, 17) \ << regs_init \ << regs_res; regs_res.write_hi_lo(true, 24); regs_res.write_hi_lo(false, 22); QTest::newRow("MTLO") << Instruction(0, 28, 0, 0, 0, 19) \ << regs_init \ << regs_res; regs_res.write_hi_lo(false, 28); QTest::newRow("MOVZ-F") << Instruction(0, 24, 24, 25, 0, 10) \ << regs_init \ << regs_res; QTest::newRow("MOVN-F") << Instruction(0, 24, 1, 25, 0, 11) \ << regs_init \ << regs_res; regs_res.write_gp(25, 55); QTest::newRow("MOVZ-T") << Instruction(0, 24, 1, 25, 0, 10) \ << regs_init \ << regs_res; QTest::newRow("MOVN-T") << Instruction(0, 24, 24, 25, 0, 11) \ << regs_init \ << regs_res; } } void MachineTests::singlecore_regs_data() { core_regs_data(); } void MachineTests::pipecore_regs_data() { core_regs_data(); } void MachineTests::singlecore_regs() { QFETCH(Instruction, i); QFETCH(Registers, init); QFETCH(Registers, res); Memory mem; // Just memory (it shouldn't be used here except instruction) mem.write_word(res.read_pc(), i.data()); // Store single instruction (anything else should be 0 so NOP effectively) Memory mem_used(mem); // Create memory copy CoreSingle core(&init, &mem_used, &mem_used, true); core.step(); // Single step should be enought as this is risc without pipeline res.pc_inc(); // We did single step so increment program counter accordingly QCOMPARE(init, res); // After doing changes from initial state this should be same state as in case of passed expected result QCOMPARE(mem, mem_used); // There should be no change in memory } void MachineTests::pipecore_regs() { QFETCH(Instruction, i); QFETCH(Registers, init); QFETCH(Registers, res); Memory mem; // Just memory (it shouldn't be used here except instruction) mem.write_word(res.read_pc(), i.data()); // Store single instruction (anything else should be 0 so NOP effectively) Memory mem_used(mem); res.pc_jmp(0x14); CorePipelined core(&init, &mem_used, &mem_used); for (int i = 0; i < 5; i++) core.step(); // Fire steps for five pipelines stages //cout << "well:" << init.read_gp(26) << ":" << regs_used.read_gp(26) << endl; QCOMPARE(init, res); // After doing changes from initial state this should be same state as in case of passed expected result QCOMPARE(mem, mem_used); // There should be no change in memory } static void core_jmp_data() { QTest::addColumn("i"); QTest::addColumn("regs"); QTest::addColumn("pc"); Registers regs; regs.write_gp(14, -22); regs.write_gp(15, 22); regs.write_gp(16, -22); regs.write_gp(12, 0x80040000); QTest::newRow("B") << Instruction(4, 0, 0, 61) \ << regs \ << regs.read_pc() + 4 + (61 << 2); QTest::newRow("BEQ") << Instruction(4, 14, 16, 61) \ << regs \ << regs.read_pc() + 4 + (61 << 2); QTest::newRow("BEQ-BACK") << Instruction(4, 14, 16, -4) \ << regs \ << regs.read_pc() + 4 - 16; QTest::newRow("BNE") << Instruction(5, 14, 15, 61) \ << regs \ << regs.read_pc() + 4 + (61 << 2); QTest::newRow("BGEZ") << Instruction(1, 15, 1, 61) \ << regs \ << regs.read_pc() + 4 + (61 << 2); QTest::newRow("BGTZ") << Instruction(7, 15, 0, 61) \ << regs \ << regs.read_pc() + 4 + (61 << 2); QTest::newRow("BLEZ") << Instruction(6, 14, 0, 61) \ << regs \ << regs.read_pc() + 4 + (61 << 2); QTest::newRow("BLTZ") << Instruction(1, 14, 0, 61) \ << regs \ << regs.read_pc() + 4 + (61 << 2); QTest::newRow("J") << Instruction(2, 24) \ << regs \ << 0x80000000 + (24 << 2); QTest::newRow("JR") << Instruction(0, 12, 0, 0, 0, 8) \ << regs \ << 0x80040000; } void MachineTests::singlecore_jmp_data() { core_jmp_data(); } void MachineTests::pipecore_jmp_data() { core_jmp_data(); } void MachineTests::singlecore_jmp() { QFETCH(Instruction, i); QFETCH(Registers, regs); QFETCH(std::uint32_t, pc); Memory mem; mem.write_word(regs.read_pc(), i.data()); Memory mem_used(mem); Registers regs_used(regs); CoreSingle core(®s_used, &mem_used, &mem_used, true); core.step(); QCOMPARE(regs.read_pc() + 4, regs_used.read_pc()); // First execute delay slot core.step(); QCOMPARE(pc, regs_used.read_pc()); // Now do jump QCOMPARE(mem, mem_used); // There should be no change in memory regs_used.pc_abs_jmp(regs.read_pc()); // Reset program counter before we do registers compare QCOMPARE(regs, regs_used); // There should be no change in registers now } void MachineTests::pipecore_jmp() { QFETCH(Instruction, i); QFETCH(Registers, regs); QFETCH(std::uint32_t, pc); Memory mem; mem.write_word(regs.read_pc(), i.data()); Memory mem_used(mem); Registers regs_used(regs); CorePipelined core(®s_used, &mem_used, &mem_used); core.step(); QCOMPARE(regs.read_pc() + 4, regs_used.read_pc()); // First just fetch core.step(); QCOMPARE(pc, regs_used.read_pc()); // Now do jump for (int i = 0; i < 3; i++) core.step(); // Follow up with three other steps to complete pipeline to be sure that instruction has no side effects QCOMPARE(mem, mem_used); // There should be no change in memory regs.pc_abs_jmp(pc + 12); // Set reference pc to three more instructions later (where regs_used should be) QCOMPARE(regs, regs_used); // There should be no change in registers now (except pc) } static void core_mem_data() { QTest::addColumn("i"); QTest::addColumn("regs_init"); QTest::addColumn("regs_res"); QTest::addColumn("mem_init"); QTest::addColumn("mem_res"); // Load { Memory mem; mem.write_word(0x24, 0xA3242526); Registers regs; regs.write_gp(1, 0x22); Registers regs_res(regs); regs_res.write_gp(21, 0xFFFFFFA3); QTest::newRow("LB") << Instruction(32, 1, 21, 0x2) \ << regs \ << regs_res \ << mem \ << mem; regs_res.write_gp(21, 0xFFFFA324); QTest::newRow("LH") << Instruction(33, 1, 21, 0x2) \ << regs \ << regs_res \ << mem \ << mem; regs_res.write_gp(21, 0xA3242526); QTest::newRow("LW") << Instruction(35, 1, 21, 0x2) \ << regs \ << regs_res \ << mem \ << mem; regs_res.write_gp(21, 0x000000A3); QTest::newRow("LBU") << Instruction(36, 1, 21, 0x2) \ << regs \ << regs_res \ << mem \ << mem; regs_res.write_gp(21, 0x0000A324); QTest::newRow("LHU") << Instruction(37, 1, 21, 0x2) \ << regs \ << regs_res \ << mem \ << mem; } // Store { Registers regs; regs.write_gp(1, 0x22); regs.write_gp(21, 0x23242526); Memory mem; mem.write_byte(0x24, 0x26); // Note: store least significant byte QTest::newRow("SB") << Instruction(40, 1, 21, 0x2) \ << regs \ << regs \ << Memory() \ << mem; mem.write_hword(0x24, 0x2526); QTest::newRow("SH") << Instruction(41, 1, 21, 0x2) \ << regs \ << regs \ << Memory() \ << mem; mem.write_word(0x24, 0x23242526); QTest::newRow("SH") << Instruction(43, 1, 21, 0x2) \ << regs \ << regs \ << Memory() \ << mem; } } void MachineTests::singlecore_mem_data() { core_mem_data(); } void MachineTests::pipecore_mem_data() { core_mem_data(); } void MachineTests::singlecore_mem() { QFETCH(Instruction, i); QFETCH(Registers, regs_init); QFETCH(Registers, regs_res); QFETCH(Memory, mem_init); QFETCH(Memory, mem_res); // Write instruction to both memories mem_init.write_word(regs_init.read_pc(), i.data()); mem_res.write_word(regs_init.read_pc(), i.data()); CoreSingle core(®s_init, &mem_init, &mem_init, true); core.step(); regs_res.pc_inc(); QCOMPARE(regs_init, regs_res); QCOMPARE(mem_init, mem_res); } void MachineTests::pipecore_mem() { QFETCH(Instruction, i); QFETCH(Registers, regs_init); QFETCH(Registers, regs_res); QFETCH(Memory, mem_init); QFETCH(Memory, mem_res); // Write instruction to both memories mem_init.write_word(regs_init.read_pc(), i.data()); mem_res.write_word(regs_init.read_pc(), i.data()); CorePipelined core(®s_init, &mem_init, &mem_init); for (int i = 0; i < 5; i++) core.step(); // Fire steps for five pipelines stages regs_res.pc_jmp(20); QCOMPARE(regs_init, regs_res); QCOMPARE(mem_init, mem_res); } /*======================================================================*/ static void core_alu_forward_data() { QTest::addColumn>("code"); QTest::addColumn("reg_init"); QTest::addColumn("reg_res"); // Note that we shouldn't be touching program counter as that is handled automatically and differs if we use pipelining // Test forwarding of ALU operands { QVector code{ // objdump -D to src: ^[^ \t]+[ \t]+([^ \t]+)[ \t]+([^ \t].*)$ 0x20020001, // addi v0,zero,1 0x20011111, // addi at,zero,4369 0x20012222, // addi at,zero,8738 0x00221820, // add v1,at,v0 0x00222020, // add a0,at,v0 0x20020003, // addi v0,zero,3 0x20011111, // addi at,zero,4369 0x20012222, // addi at,zero,8738 0x00412820, // add a1,v0,at 0x00413020, // add a2,v0,at 0x00000000, // nop }; Registers regs_init; regs_init.pc_abs_jmp(0x80020000); Registers regs_res(regs_init); regs_res.write_gp(1, 0x2222); regs_res.write_gp(2, 3); regs_res.write_gp(3, 0x2223); regs_res.write_gp(4, 0x2223); regs_res.write_gp(5, 0x2225); regs_res.write_gp(6, 0x2225); regs_res.pc_abs_jmp(regs_init.read_pc() + 4 * code.length()); QTest::newRow("alu_forward_1") << code << regs_init << regs_res; } // Test forwarding in JR and JALR { QVector code{ // start: = 0x80020000 0x3c041111, // lui a0,0x1111 0x3c052222, // lui a1,0x2222 0x0c008012, // jal 80020048 0x3c063333, // lui a2,0x3333 0x00021820, // add v1,zero,v0 0x0800800a, // j 80020028 0x20107777, // addi s0,zero, 0x7777 0x20128888, // addi s2,zero, 0x8888 0x20139999, // addi s3,zero, 0x9999 0x2014aaaa, // addi s4,zero, 0xaaaa // skip: 0x3c088002, // lui t0,0x8002 0x25080058, // addiu t0,t0,88 0x0100f809, // jalr t0 0x2004abcd, // addi a0,zero, 0xabcd 0x20040000, // addi a0,zero,0 0x20510000, // addi s1,v0,0 0x08008018, // j 80020060 0x00000000, // nop // fnc_add3: 0x00851020, // add v0,a0,a1 0x00461020, // add v0,v0,a2 0x03e00008, // jr ra 0x00000000, // nop // fnc_short: 0x03e00008, // jr ra 0x20820000, // addi v0,a0,0 // loop: 0x1000ffff, // b 80020060 0x00000000, // nop }; Registers regs_init; regs_init.pc_abs_jmp(0x80020000); Registers regs_res(regs_init); regs_res.write_gp(1, 0x00000000); regs_res.write_gp(2, 0xffffabcd); regs_res.write_gp(3, 0x66660000); regs_res.write_gp(4, 0x00000000); regs_res.write_gp(5, 0x22220000); regs_res.write_gp(6, 0x33330000); regs_res.write_gp(7, 0x00000000); regs_res.write_gp(8, 0x80020058); regs_res.write_gp(16, 0x00007777); regs_res.write_gp(17, 0xffffabcd); regs_res.write_gp(31, 0x80020038); regs_res.pc_abs_jmp(0x80020060); QTest::newRow("j_jal_jalr") << code << regs_init << regs_res; } // Test multiplication and division { QVector code{ // start: 0x3c021234, // lui v0,0x1234 0x34425678, // ori v0,v0,0x5678 0x3c03abcd, // lui v1,0xabcd 0x3463ef01, // ori v1,v1,0xef01 0x00430018, // mult v0,v1 0x00008012, // mflo s0 0x00008810, // mfhi s1 0x00430019, // multu v0,v1 0x00009012, // mflo s2 0x00009810, // mfhi s3 0x0062001a, // div zero,v1,v0 0x0000a012, // mflo s4 0x0000a810, // mfhi s5 0x0062001b, // divu zero,v1,v0 0x0000b012, // mflo s6 0x0000b810, // mfhi s7 // loop: 0x1000ffff, // b 80020070 0x00000000, // nop }; Registers regs_init; regs_init.pc_abs_jmp(0x80020000); Registers regs_res(regs_init); std::uint32_t val_a = 0x12345678; std::uint32_t val_b = 0xabcdef01; std::uint64_t val_u64; std::int64_t val_s64; regs_res.write_gp(2, val_a); regs_res.write_gp(3, val_b); val_s64 = (std::int64_t)(std::int32_t)val_a * (std::int32_t)val_b; regs_res.write_gp(16, (std::uint32_t)(val_s64 & 0xffffffff)); regs_res.write_gp(17, (std::uint32_t)(val_s64 >> 32)); val_u64 = (std::uint64_t)val_a * val_b; regs_res.write_gp(18, (std::uint32_t)(val_u64 & 0xffffffff)); regs_res.write_gp(19, (std::uint32_t)(val_u64 >> 32)); regs_res.write_gp(20, (std::uint32_t)((std::int32_t)val_b / (std::int32_t)val_a)); regs_res.write_gp(21, (std::uint32_t)((std::int32_t)val_b % (std::int32_t)val_a)); regs_res.write_gp(22, val_b / val_a); regs_res.write_gp(23, val_b % val_a); regs_res.write_hi_lo(false, regs_res.read_gp(22)); regs_res.write_hi_lo(true, regs_res.read_gp(23)); regs_res.pc_abs_jmp(regs_init.read_pc() + 4 * code.length() - 4); QTest::newRow("mul-div") << code << regs_init << regs_res; } // branches { QVector code{ // start: 0x2001ffff, // addi at,zero,-1 0x20020001, // addi v0,zero,1 0x20030000, // addi v1,zero,0 0x20040002, // addi a0,zero,2 // test_branch: 0x20050001, // addi a1,zero,1 0x04200004, // bltz at,80020028 0x00000000, // nop 0x20050000, // addi a1,zero,0 0x20631000, // addi v1,v1,4096 0x00611820, // add v1,v1,at 0x20050001, // addi a1,zero,1 0x18200004, // blez at,80020040 0x00000000, // nop 0x20050000, // addi a1,zero,0 0x20630100, // addi v1,v1,256 0x00611820, // add v1,v1,at 0x20050001, // addi a1,zero,1 0x04210004, // bgez at,80020058 0x00000000, // nop 0x20050000, // addi a1,zero,0 0x20632000, // addi v1,v1,8192 0x00611820, // add v1,v1,at 0x20050001, // addi a1,zero,1 0x1c200004, // bgtz at,80020070 0x00000000, // nop 0x20050000, // addi a1,zero,0 0x20630200, // addi v1,v1,512 0x00611820, // add v1,v1,at 0x20050001, // addi a1,zero,1 0x14220004, // bne at,v0,80020088 0x00000000, // nop 0x20050000, // addi a1,zero,0 0x20634000, // addi v1,v1,16384 0x00611820, // add v1,v1,at 0x20050001, // addi a1,zero,1 0x10220004, // beq at,v0,800200a0 0x00000000, // nop 0x20050000, // addi a1,zero,0 0x20630400, // addi v1,v1,1024 0x00611820, // add v1,v1,at 0x20210001, // addi at,at,1 0x1424ffda, // bne at,a0,80020010 0x00000000, // nop // loop: 0x1000ffff, // b 800200ac 0x00000000, // nop }; Registers regs_init; regs_init.pc_abs_jmp(0x80020000); Registers regs_res(regs_init); regs_res.write_gp(1, 2); regs_res.write_gp(2, 1); regs_res.write_gp(3, 0x8d00); regs_res.write_gp(4, 2); regs_res.write_gp(5, 1); regs_res.pc_abs_jmp(regs_init.read_pc() + 4 * code.length()); QTest::newRow("branch_conditions_test") << code << regs_init << regs_res; } } void MachineTests::singlecore_alu_forward_data() { core_alu_forward_data(); } void MachineTests::pipecore_alu_forward_data() { core_alu_forward_data(); } void MachineTests::pipecorestall_alu_forward_data() { core_alu_forward_data(); } static void run_code_fragment(Core &core, Registers ®_init, Registers ®_res, Memory &mem_init, Memory &mem_res, QVector &code) { std::uint32_t addr = reg_init.read_pc(); foreach (uint32_t i, code) { mem_init.write_word(addr, i); mem_res.write_word(addr, i); addr += 4; } for (int k = 10000; k ; k--) { core.step(); // Single step should be enought as this is risc without pipeline if (reg_init.read_pc() == reg_res.read_pc() && k > 6) // reached end of the code fragment k = 6; // add some cycles to finish processing } reg_res.pc_abs_jmp(reg_init.read_pc()); // We do not compare result pc QCOMPARE(reg_init, reg_res); // After doing changes from initial state this should be same state as in case of passed expected result QCOMPARE(mem_init, mem_res); // There should be no change in memory } void MachineTests::singlecore_alu_forward() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); Memory mem_init; Memory mem_res; CoreSingle core(®_init, &mem_init, &mem_init, true); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); } void MachineTests::pipecore_alu_forward() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); Memory mem_init; Memory mem_res; CorePipelined core(®_init, &mem_init, &mem_init, MachineConfig::HU_STALL_FORWARD); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); } void MachineTests::pipecorestall_alu_forward() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); Memory mem_init; Memory mem_res; CorePipelined core(®_init, &mem_init, &mem_init, MachineConfig::HU_STALL); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); } /*======================================================================*/ static void core_memory_tests_data() { QTest::addColumn>("code"); QTest::addColumn("reg_init"); QTest::addColumn("reg_res"); QTest::addColumn("mem_init"); QTest::addColumn("mem_res"); // Test { QVector code{ // objdump -d to src: ^[^ \t]+[ \t]+([^ \t]+)[ \t]+([^ \t].*)$ // _start: 0x20100000, // addi s0,zero,0 0x2011003c, // addi s1,zero,60 0x00109020, // add s2,zero,s0 // main_loop: 0x1211001d, // beq s0,s1,80020084 0x00000000, // nop 0x3c140000, // lui s4,0x0 0x0290a021, // addu s4,s4,s0 0x8e941000, // lw s4,4096(s4) 0x02009820, // add s3,s0,zero 0x02009020, // add s2,s0,zero // inner_loop: 0x1251000b, // beq s2,s1,80020058 0x00000000, // nop 0x3c150000, // lui s5,0x0 0x02b2a821, // addu s5,s5,s2 0x8eb51000, // lw s5,4096(s5) 0x0295082a, // slt at,s4,s5 0x14200003, // bnez at,80020050 0x00000000, // nop 0x22530000, // addi s3,s2,0 0x22b40000, // addi s4,s5,0 // not_a_min: 0x0800800a, // j 80020028 0x22520004, // addi s2,s2,4 // inner_loop_end: 0x3c150000, // lui s5,0x0 0x02b0a821, // addu s5,s5,s0 0x8eb51000, // lw s5,4096(s5) 0x3c010000, // lui at,0x0 0x00300821, // addu at,at,s0 0xac341000, // sw s4,4096(at) 0x3c010000, // lui at,0x0 0x00330821, // addu at,at,s3 0xac351000, // sw s5,4096(at) 0x08008003, // j 8002000c 0x22100004, // addi s0,s0,4 // main_loop_end: 0x20080000, // addi t0,zero,0 0xbd090000, // cache 0x9,0(t0) // end_loop: 0x08008023, // j 8002008c 0x00000000, // nop }; QVector data_init{5, 3, 4, 1, 15, 8, 9, 2, 10, 6, 11, 1, 6, 9, 12}; QVector data_res{ 1, 1, 2, 3, 4, 5, 6, 6, 8, 9, 9, 10, 11, 12, 15}; Registers regs_init; regs_init.pc_abs_jmp(0x80020000); Registers regs_res(regs_init); regs_res.write_gp(1, 0x38); regs_res.write_gp(16, 0x3c); regs_res.write_gp(17, 0x3c); regs_res.write_gp(18, 0x3c); regs_res.write_gp(19, 0x38); regs_res.write_gp(20, 0xf); regs_res.write_gp(21, 0xf); regs_res.pc_abs_jmp(regs_init.read_pc() + 4 * code.length() - 4); std::uint32_t addr; Memory mem_init; addr = 0x1000; foreach (uint32_t i, data_init) { mem_init.write_word(addr, i); addr += 4; } Memory mem_res; addr = 0x1000; foreach (uint32_t i, data_res) { mem_res.write_word(addr, i); addr += 4; } QTest::newRow("cache_insert_sort") << code << regs_init << regs_res << mem_init << mem_res; } // lwr, lwl, swr, swl { QVector code{ // __start: 0x3c1c8003, // lui gp,0x8003 0x279c90d0, // addiu gp,gp,-28464 // main: 0x3c1aaabb, // lui k0,0xaabb 0x375accdd, // ori k0,k0,0xccdd 0x23420000, // addi v0,k0,0 0x23430000, // addi v1,k0,0 0x23440000, // addi a0,k0,0 0x23450000, // addi a1,k0,0 0x23460000, // addi a2,k0,0 0x23470000, // addi a3,k0,0 0x23480000, // addi t0,k0,0 0x23490000, // addi t1,k0,0 0x234a0000, // addi t2,k0,0 0x3c1b8002, // lui k1,0x8002 0x277b0100, // addiu k1,k1,0x100 0x9b620000, // lwr v0,0(k1) 0x9b630001, // lwr v1,1(k1) 0x9b640002, // lwr a0,2(k1) 0x9b650003, // lwr a1,3(k1) 0x8b660000, // lwl a2,0(k1) 0x8b670001, // lwl a3,1(k1) 0x8b680002, // lwl t0,2(k1) 0x8b690003, // lwl t1,3(k1) 0xbb7a0000, // swr k0,0(k1) 0xbb7a0005, // swr k0,5(k1) 0xbb7a000a, // swr k0,10(k1) 0xbb7a000f, // swr k0,15(k1) 0xab7a0014, // swl k0,20(k1) 0xab7a0019, // swl k0,25(k1) 0xab7a001e, // swl k0,30(k1) 0xab7a0023, // swl k0,35(k1) 0x8f6a0000, // lw t2,0(k1) 0x8f6b0004, // lw t3,4(k1) 0x8f6c0008, // lw t4,8(k1) 0x8f6d000c, // lw t5,12(k1) 0x8f6e0010, // lw t6,16(k1) 0x8f6f0014, // lw t7,20(k1) 0x8f700018, // lw s0,24(k1) 0x8f71001c, // lw s1,28(k1) 0x8f720020, // lw s2,32(k1) 0x8f730020, // lw s3,32(k1) 0xbd090000, // cache 0x9,0(t0) // loop: 0x1000ffff, // b 800200a4 0x00000000, // nop // mem: }; Registers regs_init; regs_init.pc_abs_jmp(0x80020000); Registers regs_res(regs_init); regs_res.write_gp(2, 0xaabbcc01); regs_res.write_gp(3, 0xaabb0102); regs_res.write_gp(4, 0xaa010203); regs_res.write_gp(5, 0x01020304); regs_res.write_gp(6, 0x01020304); regs_res.write_gp(7, 0x020304dd); regs_res.write_gp(8, 0x0304ccdd); regs_res.write_gp(9, 0x04bbccdd); regs_res.write_gp(10, 0xdd020304); regs_res.write_gp(11, 0xccdd0708); regs_res.write_gp(12, 0xbbccdd0c); regs_res.write_gp(13, 0xaabbccdd); regs_res.write_gp(14, 0x11121314); regs_res.write_gp(15, 0xaabbccdd); regs_res.write_gp(16, 0x19aabbcc); regs_res.write_gp(17, 0x1d1eaabb); regs_res.write_gp(18, 0x212223aa); regs_res.write_gp(19, 0x212223aa); regs_res.write_gp(26, 0xaabbccdd); regs_res.write_gp(27, 0x80020100); regs_res.write_gp(28, 0x800290d0); std::uint32_t addr; Memory mem_init; addr = 0x80020100; QVector data_init{0x01020304, 0x05060708, 0x090a0b0c, 0x0d0e0f00, 0x11121314, 0x15161718, 0x191a1b1c, 0x1d1e1f10, 0x21222324, 0x25262728, 0x292a2b2c, 0x2d2e2f20}; foreach (uint32_t i, data_init) { mem_init.write_word(addr, i); addr += 4; } Memory mem_res; addr = 0x80020100; QVector data_res{0xdd020304, 0xccdd0708, 0xbbccdd0c, 0xaabbccdd, 0x11121314, 0xaabbccdd, 0x19aabbcc, 0x1d1eaabb, 0x212223aa, 0x25262728, 0x292a2b2c, 0x2d2e2f20}; foreach (uint32_t i, data_res) { mem_res.write_word(addr, i); addr += 4; } regs_res.pc_abs_jmp(regs_init.read_pc() + 4 * code.length() - 4); QTest::newRow("lwr_lrl_swr_swl") << code << regs_init << regs_res << mem_init << mem_res; } } void MachineTests::singlecore_memory_tests_data() { core_memory_tests_data(); } void MachineTests::pipecore_nc_memory_tests_data() { core_memory_tests_data(); } void MachineTests::pipecore_wt_na_memory_tests_data() { core_memory_tests_data(); } void MachineTests::pipecore_wt_a_memory_tests_data() { core_memory_tests_data(); } void MachineTests::pipecore_wb_memory_tests_data() { core_memory_tests_data(); } void MachineTests::singlecore_memory_tests() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); QFETCH(Memory, mem_init); QFETCH(Memory, mem_res); CoreSingle core(®_init, &mem_init, &mem_init, true); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); } void MachineTests::pipecore_nc_memory_tests() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); QFETCH(Memory, mem_init); QFETCH(Memory, mem_res); CorePipelined core(®_init, &mem_init, &mem_init, MachineConfig::HU_STALL_FORWARD); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); } void MachineTests::pipecore_wt_na_memory_tests() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); QFETCH(Memory, mem_init); QFETCH(Memory, mem_res); MachineConfigCache cache_conf; cache_conf.set_enabled(true); cache_conf.set_sets(2); // Number of sets cache_conf.set_blocks(1); // Number of blocks cache_conf.set_associativity(2); // Degree of associativity cache_conf.set_replacement_policy(MachineConfigCache::RP_LRU); cache_conf.set_write_policy(MachineConfigCache::WP_TROUGH_NOALLOC); Cache i_cache(&mem_init, &cache_conf); Cache d_cache(&mem_init, &cache_conf); CorePipelined core(®_init, &i_cache, &d_cache, MachineConfig::HU_STALL_FORWARD); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); } void MachineTests::pipecore_wt_a_memory_tests() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); QFETCH(Memory, mem_init); QFETCH(Memory, mem_res); MachineConfigCache cache_conf; cache_conf.set_enabled(true); cache_conf.set_sets(2); // Number of sets cache_conf.set_blocks(1); // Number of blocks cache_conf.set_associativity(2); // Degree of associativity cache_conf.set_replacement_policy(MachineConfigCache::RP_LRU); cache_conf.set_write_policy(MachineConfigCache::WP_TROUGH_ALLOC); Cache i_cache(&mem_init, &cache_conf); Cache d_cache(&mem_init, &cache_conf); CorePipelined core(®_init, &i_cache, &d_cache, MachineConfig::HU_STALL_FORWARD); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); } void MachineTests::pipecore_wb_memory_tests() { QFETCH(QVector, code); QFETCH(Registers, reg_init); QFETCH(Registers, reg_res); QFETCH(Memory, mem_init); QFETCH(Memory, mem_res); MachineConfigCache cache_conf; cache_conf.set_enabled(true); cache_conf.set_sets(4); // Number of sets cache_conf.set_blocks(2); // Number of blocks cache_conf.set_associativity(2); // Degree of associativity cache_conf.set_replacement_policy(MachineConfigCache::RP_LRU); cache_conf.set_write_policy(MachineConfigCache::WP_BACK); Cache i_cache(&mem_init, &cache_conf); Cache d_cache(&mem_init, &cache_conf); CorePipelined core(®_init, &i_cache, &d_cache, MachineConfig::HU_STALL_FORWARD); run_code_fragment(core, reg_init, reg_res, mem_init, mem_res, code); }