/* * Copyright © 2020 Valve Corporation * * SPDX-License-Identifier: MIT */ #include #include "helpers.h" #include "sid.h" using namespace aco; static std::vector filter_gfx_levels(std::vector src) { std::vector res; for (amd_gfx_level gfx : src) { if (gfx < GFX12 || LLVM_VERSION_MAJOR >= 19) res.push_back(gfx); } return res; } BEGIN_TEST(assembler.s_memtime) for (unsigned i = GFX6; i <= GFX10; i++) { if (!setup_cs(NULL, (amd_gfx_level)i)) continue; //~gfx[6-7]>> c7800000 //~gfx[6-7]! bf810000 //~gfx[8-9]>> s_memtime s[0:1] ; c0900000 00000000 //~gfx10>> s_memtime s[0:1] ; f4900000 fa000000 bld.smem(aco_opcode::s_memtime, bld.def(s2)).def(0).setFixed(PhysReg{0}); finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.branch_3f) if (!setup_cs(NULL, (amd_gfx_level)GFX10)) return; //! BB0: //! s_branch BB1 ; bf820040 //! s_nop 0 ; bf800000 bld.sopp(aco_opcode::s_branch, 1); for (unsigned i = 0; i < 0x3f; i++) bld.vop1(aco_opcode::v_nop); bld.reset(program->create_and_insert_block()); program->blocks[1].linear_preds.push_back(0u); finish_assembler_test(); END_TEST BEGIN_TEST(assembler.long_jump.unconditional_forwards) if (!setup_cs(NULL, (amd_gfx_level)GFX10)) return; //!BB0: //! s_branch 16369 ; bf823ff1 bld.sopp(aco_opcode::s_branch, 2); bld.reset(program->create_and_insert_block()); //! s_nop 0 ; bf800000 //!(then repeated 16366 times) for (unsigned i = 0; i < INT16_MAX + 1; i++) bld.sopp(aco_opcode::s_nop, 0); //! s_waitcnt_vscnt null, 0x0 ; bbfd0000 //! s_branch 1 ; bf820001 //! s_branch BB2 ; bf824011 //! s_nop 0 ; bf800000 //! (then repeated 16400 times) //! BB2: //! s_endpgm ; bf810000 bld.reset(program->create_and_insert_block()); program->blocks[2].linear_preds.push_back(0u); program->blocks[2].linear_preds.push_back(1u); finish_assembler_test(); END_TEST BEGIN_TEST(assembler.long_jump.conditional_forwards) for (amd_gfx_level gfx : filter_gfx_levels({GFX10, GFX12})) { if (!setup_cs(NULL, gfx)) continue; //! BB0: //~gfx10! s_cbranch_scc0 16369 ; bf843ff1 //~gfx12! s_cbranch_scc0 16368 ; bfa13ff0 bld.sopp(aco_opcode::s_cbranch_scc0, 2); bld.reset(program->create_and_insert_block()); //! BB1: //! s_nop 0 ; bf800000 //!(then repeated 16366 times) //~gfx10! s_waitcnt_vscnt null, 0x0 ; bbfd0000 //! s_branch 1 ; $_ //! s_branch BB2 ; $_ //! s_nop 0 ; bf800000 //!(then repeated 16400 times) for (unsigned i = 0; i < INT16_MAX + 1; i++) bld.sopp(aco_opcode::s_nop, 0); //! BB2: //! s_endpgm ; $_ bld.reset(program->create_and_insert_block()); program->blocks[1].linear_preds.push_back(0u); program->blocks[2].linear_preds.push_back(0u); program->blocks[2].linear_preds.push_back(1u); finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.long_jump.unconditional_backwards) if (!setup_cs(NULL, (amd_gfx_level)GFX10)) return; //!BB0: //! s_nop 0 ; bf800000 //!(then repeated 16367 times) for (unsigned i = 0; i < INT16_MAX + 1; i++) bld.sopp(aco_opcode::s_nop, 0); //! s_waitcnt_vscnt null, 0x0 ; bbfd0000 //! s_branch 1 ; bf820001 //! s_branch BB0 ; bf82c00d //! s_nop 0 ; bf800000 //! (then repeated 16399 times) //! s_branch 49134 ; bf82bfee bld.sopp(aco_opcode::s_branch, 0); //! BB1: //! s_endpgm ; bf810000 bld.reset(program->create_and_insert_block()); program->blocks[0].linear_preds.push_back(0u); program->blocks[1].linear_preds.push_back(0u); finish_assembler_test(); END_TEST BEGIN_TEST(assembler.long_jump.conditional_backwards) if (!setup_cs(NULL, (amd_gfx_level)GFX10)) return; //!BB0: //! s_nop 0 ; bf800000 //!(then repeated 16367 times) for (unsigned i = 0; i < INT16_MAX + 1; i++) bld.sopp(aco_opcode::s_nop, 0); //! s_waitcnt_vscnt null, 0x0 ; bbfd0000 //! s_branch 1 ; bf820001 //! s_branch BB0 ; bf82c00d //! s_nop 0 ; bf800000 //!(then repeated 16399 times) //! s_cbranch_execnz 49134 ; bf89bfee bld.sopp(aco_opcode::s_cbranch_execnz, 0); //! BB1: //! s_endpgm ; bf810000 bld.reset(program->create_and_insert_block()); program->blocks[0].linear_preds.push_back(0u); program->blocks[1].linear_preds.push_back(0u); finish_assembler_test(); END_TEST BEGIN_TEST(assembler.long_jump.constaddr) if (!setup_cs(NULL, (amd_gfx_level)GFX10)) return; //>> s_branch 16369 ; bf823ff1 bld.sopp(aco_opcode::s_branch, 2); bld.reset(program->create_and_insert_block()); for (unsigned i = 0; i < INT16_MAX + 1; i++) bld.sopp(aco_opcode::s_nop, 0); bld.reset(program->create_and_insert_block()); //>> s_getpc_b64 s[0:1] ; be801f00 //! s_add_u32 s0, s0, 32 ; 8000ff00 00000020 bld.sop1(aco_opcode::p_constaddr_getpc, Definition(PhysReg(0), s2), Operand::zero()); bld.sop2(aco_opcode::p_constaddr_addlo, Definition(PhysReg(0), s1), bld.def(s1, scc), Operand(PhysReg(0), s1), Operand::zero(), Operand::zero()); program->blocks[2].linear_preds.push_back(0u); program->blocks[2].linear_preds.push_back(1u); finish_assembler_test(); END_TEST BEGIN_TEST(assembler.long_jump.discard_early_exit) if (!setup_cs(NULL, (amd_gfx_level)GFX10)) return; //! BB0: //! s_cbranch_scc0 16369 ; bf843ff1 bld.sopp(aco_opcode::s_cbranch_scc0, 2); bld.reset(program->create_and_insert_block()); //! BB1: //! s_nop 1 ; bf800001 //! (then repeated 16366 times) //! s_waitcnt_vscnt null, 0x0 ; bbfd0000 //! s_branch 1 ; bf820001 //! s_branch BB2 ; bf824011 //! s_nop 1 ; bf800001 //! (then repeated 16399 times) //! s_endpgm ; bf810000 for (unsigned i = 0; i < INT16_MAX; i++) bld.sopp(aco_opcode::s_nop, 1); //! BB2: //! s_endpgm ; bf810000 bld.reset(program->create_and_insert_block()); program->blocks[1].linear_preds.push_back(0u); program->blocks[2].linear_preds.push_back(0u); program->blocks[2].kind = block_kind_discard_early_exit; finish_assembler_test(); END_TEST BEGIN_TEST(assembler.v_add3) for (unsigned i = GFX9; i <= GFX10; i++) { if (!setup_cs(NULL, (amd_gfx_level)i)) continue; //~gfx9>> v_add3_u32 v0, 0, 0, 0 ; d1ff0000 02010080 //~gfx10>> v_add3_u32 v0, 0, 0, 0 ; d76d0000 02010080 aco_ptr add3{create_instruction(aco_opcode::v_add3_u32, Format::VOP3, 3, 1)}; add3->operands[0] = Operand::zero(); add3->operands[1] = Operand::zero(); add3->operands[2] = Operand::zero(); add3->definitions[0] = Definition(PhysReg(0), v1); bld.insert(std::move(add3)); finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.v_add3_clamp) for (unsigned i = GFX9; i <= GFX10; i++) { if (!setup_cs(NULL, (amd_gfx_level)i)) continue; //~gfx9>> integer addition + clamp ; d1ff8000 02010080 //~gfx10>> integer addition + clamp ; d76d8000 02010080 aco_ptr add3{create_instruction(aco_opcode::v_add3_u32, Format::VOP3, 3, 1)}; add3->operands[0] = Operand::zero(); add3->operands[1] = Operand::zero(); add3->operands[2] = Operand::zero(); add3->definitions[0] = Definition(PhysReg(0), v1); add3->valu().clamp = 1; bld.insert(std::move(add3)); finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.smem_offset) for (unsigned i = GFX9; i <= GFX10; i++) { if (!setup_cs(NULL, (amd_gfx_level)i)) continue; Definition dst(PhysReg(7), s1); Operand sbase(PhysReg(6), s2); Operand offset(PhysReg(5), s1); //~gfx9>> s_load_dword s7, s[6:7], s5 ; c00001c3 00000005 //~gfx10>> s_load_dword s7, s[6:7], s5 ; f40001c3 0a000000 bld.smem(aco_opcode::s_load_dword, dst, sbase, offset); //~gfx9! s_load_dword s7, s[6:7], 0x42 ; c00201c3 00000042 //~gfx10! s_load_dword s7, s[6:7], 0x42 ; f40001c3 fa000042 bld.smem(aco_opcode::s_load_dword, dst, sbase, Operand::c32(0x42)); if (i >= GFX9) { //~gfx9! s_load_dword s7, s[6:7], s5 offset:0x42 ; c00241c3 0a000042 //~gfx10! s_load_dword s7, s[6:7], s5 offset:0x42 ; f40001c3 0a000042 bld.smem(aco_opcode::s_load_dword, dst, sbase, Operand::c32(0x42), offset); } finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.p_constaddr) if (!setup_cs(NULL, GFX9)) return; Definition dst0 = bld.def(s2); Definition dst1 = bld.def(s2); dst0.setFixed(PhysReg(0)); dst1.setFixed(PhysReg(2)); //>> s_getpc_b64 s[0:1] ; be801c00 //! s_add_u32 s0, s0, 44 ; 8000ff00 0000002c bld.pseudo(aco_opcode::p_constaddr, dst0, bld.def(s1, scc), Operand::zero()); //! s_getpc_b64 s[2:3] ; be821c00 //! s_add_u32 s2, s2, 64 ; 8002ff02 00000040 bld.pseudo(aco_opcode::p_constaddr, dst1, bld.def(s1, scc), Operand::c32(32)); aco::lower_to_hw_instr(program.get()); finish_assembler_test(); END_TEST BEGIN_TEST(assembler.vopc_sdwa) for (unsigned i = GFX9; i <= GFX10; i++) { if (!setup_cs(NULL, (amd_gfx_level)i)) continue; //~gfx9>> v_cmp_lt_u32_sdwa vcc, 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7d9300f9 86860080 //~gfx10>> v_cmp_lt_u32_sdwa vcc, 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7d8300f9 86860080 bld.vopc_sdwa(aco_opcode::v_cmp_lt_u32, Definition(vcc, s2), Operand::zero(), Operand::zero()); //~gfx9! v_cmp_lt_u32_sdwa s[44:45], 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7d9300f9 8686ac80 //~gfx10! v_cmp_lt_u32_sdwa s[44:45], 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7d8300f9 8686ac80 bld.vopc_sdwa(aco_opcode::v_cmp_lt_u32, Definition(PhysReg(0x2c), s2), Operand::zero(), Operand::zero()); //~gfx9! v_cmp_lt_u32_sdwa exec, 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7d9300f9 8686fe80 //~gfx10! v_cmp_lt_u32_sdwa exec, 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7d8300f9 8686fe80 bld.vopc_sdwa(aco_opcode::v_cmp_lt_u32, Definition(exec, s2), Operand::zero(), Operand::zero()); if (i == GFX10) { //~gfx10! v_cmpx_lt_u32_sdwa 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7da300f9 86860080 bld.vopc_sdwa(aco_opcode::v_cmpx_lt_u32, Definition(exec, s2), Operand::zero(), Operand::zero()); } else { //~gfx9! v_cmpx_lt_u32_sdwa vcc, 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7db300f9 86860080 bld.vopc_sdwa(aco_opcode::v_cmpx_lt_u32, Definition(vcc, s2), Definition(exec, s2), Operand::zero(), Operand::zero()); //~gfx9! v_cmpx_lt_u32_sdwa s[44:45], 0, 0 src0_sel:DWORD src1_sel:DWORD ; 7db300f9 8686ac80 bld.vopc_sdwa(aco_opcode::v_cmpx_lt_u32, Definition(PhysReg(0x2c), s2), Definition(exec, s2), Operand::zero(), Operand::zero()); } finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.smem) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst = bld.def(s1); dst.setFixed(PhysReg(4)); Operand op_s1(bld.tmp(s1)); op_s1.setFixed(PhysReg(8)); Operand op_s2(bld.tmp(s2)); op_s2.setFixed(PhysReg(16)); Operand op_s4(bld.tmp(s4)); op_s4.setFixed(PhysReg(32)); //~gfx11>> s_dcache_inv ; f4840000 f8000000 //~gfx12>> s_dcache_inv ; f4042000 f8000000 bld.smem(aco_opcode::s_dcache_inv); //! s_load_b32 s4, s[16:17], 0x2a ; f4000108 f800002a bld.smem(aco_opcode::s_load_dword, dst, op_s2, Operand::c32(42)); //~gfx11! s_load_b32 s4, s[16:17], s8 ; f4000108 10000000 //~gfx12! s_load_b32 s4, s[16:17], s8 offset:0x0 ; f4000108 10000000 bld.smem(aco_opcode::s_load_dword, dst, op_s2, op_s1); //! s_load_b32 s4, s[16:17], s8 offset:0x2a ; f4000108 1000002a bld.smem(aco_opcode::s_load_dword, dst, op_s2, Operand::c32(42), op_s1); ac_hw_cache_flags cache_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_non_temporal = {{0, 0, 0, 0, 0}}; if (gfx >= GFX12) { cache_coherent.gfx12.scope = gfx12_scope_device; cache_non_temporal.gfx12.temporal_hint = gfx12_load_non_temporal; } else { cache_coherent.value = ac_glc; cache_non_temporal.value = ac_dlc; } //~gfx11! s_buffer_load_b32 s4, s[32:35], s8 glc ; f4204110 10000000 //~gfx12! s_buffer_load_b32 s4, s[32:35], s8 offset:0x0 scope:SCOPE_DEV ; f4420110 10000000 bld.smem(aco_opcode::s_buffer_load_dword, dst, op_s4, op_s1)->smem().cache = cache_coherent; //~gfx11! s_buffer_load_b32 s4, s[32:35], s8 dlc ; f4202110 10000000 //~gfx12! s_buffer_load_b32 s4, s[32:35], s8 offset:0x0 th:TH_LOAD_NT ; f4820110 10000000 bld.smem(aco_opcode::s_buffer_load_dword, dst, op_s4, op_s1)->smem().cache = cache_non_temporal; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.mubuf) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst = bld.def(v1); dst.setFixed(PhysReg(256 + 42)); Operand op_s4(bld.tmp(s4)); op_s4.setFixed(PhysReg(32)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 10)); Operand op_v2(bld.tmp(v2)); op_v2.setFixed(PhysReg(256 + 20)); Operand op_s1(bld.tmp(s1)); op_s1.setFixed(PhysReg(30)); Operand op_m0(bld.tmp(s1)); op_m0.setFixed(m0); //! llvm_version: #llvm_ver fprintf(output, "llvm_version: %u\n", LLVM_VERSION_MAJOR); /* Addressing */ //~gfx11>> buffer_load_b32 v42, off, s[32:35], s30 ; e0500000 1e082a80 //~gfx12>> buffer_load_b32 v42, off, s[32:35], s30 ; c405001e 0080402a 00000000 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), op_s1, 0, false); //~gfx11! buffer_load_b32 v42, off, s[32:35], 0 ; e0500000 80082a80 //~gfx12! buffer_load_b32 v42, off, s[32:35], null ; c405007c 0080402a 00000000 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), Operand::zero(), 0, false); //~gfx11! buffer_load_b32 v42, off, s[32:35], 42 ; e0500000 aa082a80 if (gfx == GFX11) bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), Operand::c32(42), 0, false); //~gfx11! buffer_load_b32 v42, v10, s[32:35], s30 offen ; e0500000 1e482a0a //~gfx12! buffer_load_b32 v42, v10, s[32:35], s30 offen ; c405001e 4080402a 0000000a bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, op_v1, op_s1, 0, true); //~gfx11! buffer_load_b32 v42, v10, s[32:35], s30 idxen ; e0500000 1e882a0a //~gfx12! buffer_load_b32 v42, v10, s[32:35], s30 idxen ; c405001e 8080402a 0000000a bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, op_v1, op_s1, 0, false)->mubuf().idxen = true; //~gfx11! buffer_load_b32 v42, v[20:21], s[32:35], s30 idxen offen ; e0500000 1ec82a14 //~gfx12! buffer_load_b32 v42, v[20:21], s[32:35], s30 idxen offen ; c405001e c080402a 00000014 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, op_v2, op_s1, 0, true)->mubuf().idxen = true; //~gfx11! buffer_load_b32 v42, off, s[32:35], s30 offset:84 ; e0500054 1e082a80 //~gfx12! buffer_load_b32 v42, off, s[32:35], s30 offset:84 ; c405001e 0080402a 00005400 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), op_s1, 84, false); /* Various flags */ ac_hw_cache_flags cache_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_sys_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_non_temporal = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_atomic_rtn = {{0, 0, 0, 0, 0}}; if (gfx >= GFX12) { cache_coherent.gfx12.scope = gfx12_scope_device; cache_sys_coherent.gfx12.scope = gfx12_scope_memory; cache_non_temporal.gfx12.temporal_hint = gfx12_load_non_temporal; cache_atomic_rtn.gfx12.temporal_hint = gfx12_atomic_return; } else { cache_coherent.value = ac_glc; cache_sys_coherent.value = ac_slc; cache_non_temporal.value = ac_dlc; cache_atomic_rtn.value = ac_glc; } //~gfx11! buffer_load_b32 v42, off, s[32:35], 0 glc ; e0504000 80082a80 //~gfx12! buffer_load_b32 v42, off, s[32:35], null scope:SCOPE_DEV ; c405007c 0088402a 00000000 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), Operand::zero(), 0, false) ->mubuf() .cache = cache_coherent; //~gfx11! buffer_load_b32 v42, off, s[32:35], 0 dlc ; e0502000 80082a80 //~gfx12! buffer_load_b32 v42, off, s[32:35], null th:TH_LOAD_NT ; c405007c 0090402a 00000000 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), Operand::zero(), 0, false) ->mubuf() .cache = cache_non_temporal; //~gfx11! buffer_load_b32 v42, off, s[32:35], 0 slc ; e0501000 80082a80 //~gfx12! buffer_load_b32 v42, off, s[32:35], null scope:SCOPE_SYS ; c405007c 008c402a 00000000 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), Operand::zero(), 0, false) ->mubuf() .cache = cache_sys_coherent; //; if llvm_ver >= 16 and variant == 'gfx11': //; insert_pattern('buffer_load_b32 v[42:43], off, s[32:35], 0 tfe ; e0500000 80282a80') //; elif variant == 'gfx11': //; insert_pattern('buffer_load_b32 v42, off, s[32:35], 0 tfe ; e0500000 80282a80') //~gfx12! buffer_load_b32 v[42:43], off, s[32:35], null tfe ; c445007c 0080402a 00000000 bld.mubuf(aco_opcode::buffer_load_dword, dst, op_s4, Operand(v1), Operand::zero(), 0, false) ->mubuf() .tfe = true; /* LDS */ if (gfx == GFX11) { //~gfx11! buffer_load_lds_b32 off, s[32:35], 0 ; e0c40000 80080080 bld.mubuf(aco_opcode::buffer_load_dword, op_s4, Operand(v1), Operand::zero(), op_m0, 0, false) ->mubuf() .lds = true; //~gfx11! buffer_load_lds_i8 off, s[32:35], 0 ; e0b80000 80080080 bld.mubuf(aco_opcode::buffer_load_sbyte, op_s4, Operand(v1), Operand::zero(), op_m0, 0, false) ->mubuf() .lds = true; //~gfx11! buffer_load_lds_i16 off, s[32:35], 0 ; e0c00000 80080080 bld.mubuf(aco_opcode::buffer_load_sshort, op_s4, Operand(v1), Operand::zero(), op_m0, 0, false) ->mubuf() .lds = true; //~gfx11! buffer_load_lds_u8 off, s[32:35], 0 ; e0b40000 80080080 bld.mubuf(aco_opcode::buffer_load_ubyte, op_s4, Operand(v1), Operand::zero(), op_m0, 0, false) ->mubuf() .lds = true; //~gfx11! buffer_load_lds_u16 off, s[32:35], 0 ; e0bc0000 80080080 bld.mubuf(aco_opcode::buffer_load_ushort, op_s4, Operand(v1), Operand::zero(), op_m0, 0, false) ->mubuf() .lds = true; //~gfx11! buffer_load_lds_format_x off, s[32:35], 0 ; e0c80000 80080080 bld.mubuf(aco_opcode::buffer_load_format_x, op_s4, Operand(v1), Operand::zero(), op_m0, 0, false) ->mubuf() .lds = true; } /* Stores */ //~gfx11! buffer_store_b32 v10, off, s[32:35], s30 ; e0680000 1e080a80 //~gfx12! buffer_store_b32 v10, off, s[32:35], s30 ; c406801e 0080400a 00000000 bld.mubuf(aco_opcode::buffer_store_dword, op_s4, Operand(v1), op_s1, op_v1, 0, false); //~gfx11! buffer_store_b64 v[20:21], v10, s[32:35], s30 offen ; e06c0000 1e48140a //~gfx12! buffer_store_b64 v[20:21], v10, s[32:35], s30 offen ; c406c01e 40804014 0000000a bld.mubuf(aco_opcode::buffer_store_dwordx2, op_s4, op_v1, op_s1, op_v2, 0, true); /* Atomic with return */ //~gfx11! buffer_atomic_add_u32 v10, off, s[32:35], 0 glc ; e0d44000 80080a80 //~gfx12! buffer_atomic_add_u32 v10, off, s[32:35], null th:TH_ATOMIC_RETURN ; c40d407c 0090400a 00000000 bld.mubuf(aco_opcode::buffer_atomic_add, Definition(op_v1.physReg(), v1), op_s4, Operand(v1), Operand::zero(), op_v1, 0, false) ->mubuf() .cache = cache_atomic_rtn; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.mtbuf) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst = bld.def(v1); dst.setFixed(PhysReg(256 + 42)); Operand op_s4(bld.tmp(s4)); op_s4.setFixed(PhysReg(32)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 10)); Operand op_v2(bld.tmp(v2)); op_v2.setFixed(PhysReg(256 + 20)); Operand op_s1(bld.tmp(s1)); op_s1.setFixed(PhysReg(30)); unsigned dfmt = V_008F0C_BUF_DATA_FORMAT_32_32; unsigned nfmt = V_008F0C_BUF_NUM_FORMAT_FLOAT; //! llvm_version: #llvm_ver fprintf(output, "llvm_version: %u\n", LLVM_VERSION_MAJOR); /* Addressing */ //~gfx11>> tbuffer_load_format_x v42, off, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] ; e9900000 1e082a80 //~gfx12>> tbuffer_load_format_x v42, off, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] ; c420001e 1900402a 00000080 bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), op_s1, dfmt, nfmt, 0, false); //~gfx11! tbuffer_load_format_x v42, off, s[32:35], 0 format:[BUF_FMT_32_32_FLOAT] ; e9900000 80082a80 //~gfx12! tbuffer_load_format_x v42, off, s[32:35], null format:[BUF_FMT_32_32_FLOAT] ; c420007c 1900402a 00000080 bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), Operand::zero(), dfmt, nfmt, 0, false); //~gfx11! tbuffer_load_format_x v42, off, s[32:35], 42 format:[BUF_FMT_32_32_FLOAT] ; e9900000 aa082a80 if (gfx == GFX11) bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), Operand::c32(42), dfmt, nfmt, 0, false); //~gfx11! tbuffer_load_format_x v42, v10, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] offen ; e9900000 1e482a0a //~gfx12! tbuffer_load_format_x v42, v10, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] offen ; c420001e 5900402a 0000000a bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, op_v1, op_s1, dfmt, nfmt, 0, true); //~gfx11! tbuffer_load_format_x v42, v10, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] idxen ; e9900000 1e882a0a //~gfx12! tbuffer_load_format_x v42, v10, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] idxen ; c420001e 9900402a 0000000a bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, op_v1, op_s1, dfmt, nfmt, 0, false) ->mtbuf() .idxen = true; //~gfx11! tbuffer_load_format_x v42, v[20:21], s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] idxen offen ; e9900000 1ec82a14 //~gfx12! tbuffer_load_format_x v42, v[20:21], s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] idxen offen ; c420001e d900402a 00000014 bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, op_v2, op_s1, dfmt, nfmt, 0, true) ->mtbuf() .idxen = true; //~gfx11! tbuffer_load_format_x v42, off, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] offset:84 ; e9900054 1e082a80 //~gfx12! tbuffer_load_format_x v42, off, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] offset:84 ; c420001e 1900402a 00005480 bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), op_s1, dfmt, nfmt, 84, false); /* Various flags */ ac_hw_cache_flags cache_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_sys_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_non_temporal = {{0, 0, 0, 0, 0}}; if (gfx >= GFX12) { cache_coherent.gfx12.scope = gfx12_scope_device; cache_sys_coherent.gfx12.scope = gfx12_scope_memory; cache_non_temporal.gfx12.temporal_hint = gfx12_load_non_temporal; } else { cache_coherent.value = ac_glc; cache_sys_coherent.value = ac_slc; cache_non_temporal.value = ac_dlc; } //~gfx11! tbuffer_load_format_x v42, off, s[32:35], 0 format:[BUF_FMT_32_32_FLOAT] glc ; e9904000 80082a80 //~gfx12! tbuffer_load_format_x v42, off, s[32:35], null format:[BUF_FMT_32_32_FLOAT] scope:SCOPE_DEV ; c420007c 1908402a 00000080 bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), Operand::zero(), dfmt, nfmt, 0, false) ->mtbuf() .cache = cache_coherent; //~gfx11! tbuffer_load_format_x v42, off, s[32:35], 0 format:[BUF_FMT_32_32_FLOAT] dlc ; e9902000 80082a80 //~gfx12! tbuffer_load_format_x v42, off, s[32:35], null format:[BUF_FMT_32_32_FLOAT] th:TH_LOAD_NT ; c420007c 1910402a 00000080 bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), Operand::zero(), dfmt, nfmt, 0, false) ->mtbuf() .cache = cache_non_temporal; //~gfx11! tbuffer_load_format_x v42, off, s[32:35], 0 format:[BUF_FMT_32_32_FLOAT] slc ; e9901000 80082a80 //~gfx12! tbuffer_load_format_x v42, off, s[32:35], null format:[BUF_FMT_32_32_FLOAT] scope:SCOPE_SYS ; c420007c 190c402a 00000080 bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), Operand::zero(), dfmt, nfmt, 0, false) ->mtbuf() .cache = cache_sys_coherent; //; if llvm_ver >= 19 and variant == 'gfx11': //; insert_pattern('(invalid instruction) ; e9900000') //; insert_pattern('s_add_u32 s40, 0, s42 ; 80282a80') //; elif llvm_ver >= 19 and variant == 'gfx12': //; insert_pattern('(invalid instruction) ; c460007c') //; insert_pattern('v_mul_hi_u32_u24_e32 v128, s42, v32 ; 1900402a') //; insert_pattern('(invalid instruction) ; 00000080') //; elif llvm_ver >= 16 and variant == 'gfx11': //; insert_pattern('tbuffer_load_format_x v42, off, s[32:35], 0 format:[BUF_FMT_32_32_FLOAT] ; e9900000 80282a80') //; elif variant == 'gfx11': //; insert_pattern('tbuffer_load_format_x v42, off, s[32:35], 0 format:[BUF_FMT_32_32_FLOAT] tfe ; e9900000 80282a80') //; elif variant == 'gfx12': //; insert_pattern('tbuffer_load_format_x v42, off, s[32:35], null format:[BUF_FMT_32_32_FLOAT] ; c460007c 1900402a 00000080') bld.mtbuf(aco_opcode::tbuffer_load_format_x, dst, op_s4, Operand(v1), Operand::zero(), dfmt, nfmt, 0, false) ->mtbuf() .tfe = true; /* Stores */ //~gfx11! tbuffer_store_format_x v10, off, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] ; e9920000 1e080a80 //~gfx12! tbuffer_store_format_x v10, off, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] ; c421001e 1900400a 00000080 bld.mtbuf(aco_opcode::tbuffer_store_format_x, op_s4, Operand(v1), op_s1, op_v1, dfmt, nfmt, 0, false); //~gfx11! tbuffer_store_format_xy v[20:21], v10, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] offen ; e9928000 1e48140a //~gfx12! tbuffer_store_format_xy v[20:21], v10, s[32:35], s30 format:[BUF_FMT_32_32_FLOAT] offen ; c421401e 59004014 0000000a bld.mtbuf(aco_opcode::tbuffer_store_format_xy, op_s4, op_v1, op_s1, op_v2, dfmt, nfmt, 0, true); finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.mimg) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst_v1 = bld.def(v1); dst_v1.setFixed(PhysReg(256 + 42)); Definition dst_v4 = bld.def(v4); dst_v4.setFixed(PhysReg(256 + 84)); Operand op_s4(bld.tmp(s4)); op_s4.setFixed(PhysReg(32)); Operand op_s8(bld.tmp(s8)); op_s8.setFixed(PhysReg(64)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 10)); Operand op_v2(bld.tmp(v2)); op_v2.setFixed(PhysReg(256 + 20)); Operand op_v4(bld.tmp(v4)); op_v4.setFixed(PhysReg(256 + 30)); //~gfx11>> image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D ; f06c0f00 2010540a //~gfx12>> image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D ; e7c6c000 10008054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1); //~gfx11! image_sample v[84:87], v[20:21], s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_2D ; f06c0f04 20105414 //~gfx12! image_sample v[84:87], [v20, v21], s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_2D ; e7c6c001 10008054 00001514 bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v2)->mimg().dim = ac_image_2d; //~gfx11! image_sample v42, v10, s[64:71], s[32:35] dmask:0x1 dim:SQ_RSRC_IMG_1D ; f06c0100 20102a0a //~gfx12! image_sample v42, v10, s[64:71], s[32:35] dmask:0x1 dim:SQ_RSRC_IMG_1D ; e446c000 1000802a 0000000a bld.mimg(aco_opcode::image_sample, dst_v1, op_s8, op_s4, Operand(v1), op_v1)->mimg().dmask = 0x1; /* Various flags */ ac_hw_cache_flags cache_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_sys_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_non_temporal = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_atomic_rtn = {{0, 0, 0, 0, 0}}; if (gfx >= GFX12) { cache_coherent.gfx12.scope = gfx12_scope_device; cache_sys_coherent.gfx12.scope = gfx12_scope_memory; cache_non_temporal.gfx12.temporal_hint = gfx12_load_non_temporal; cache_atomic_rtn.gfx12.temporal_hint = gfx12_atomic_return; } else { cache_coherent.value = ac_glc; cache_sys_coherent.value = ac_slc; cache_non_temporal.value = ac_dlc; cache_atomic_rtn.value = ac_glc; } //~gfx11! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D dlc ; f06c2f00 2010540a //~gfx12! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D th:TH_LOAD_NT ; e7c6c000 10108054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().cache = cache_non_temporal; //~gfx11! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D glc ; f06c4f00 2010540a //~gfx12! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D scope:SCOPE_DEV ; e7c6c000 10088054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().cache = cache_coherent; //~gfx11! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D slc ; f06c1f00 2010540a //~gfx12! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D scope:SCOPE_SYS ; e7c6c000 100c8054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().cache = cache_sys_coherent; //~gfx11! image_sample v[84:88], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D tfe ; f06c0f00 2030540a //~gfx12! image_sample v[84:88], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D tfe ; e7c6c008 10008054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().tfe = true; //~gfx11! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D lwe ; f06c0f00 2050540a //~gfx12! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D lwe ; e7c6c000 10008154 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().lwe = true; //~gfx11! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D r128 ; f06c8f00 2010540a //~gfx12! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D r128 ; e7c6c010 10008054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().r128 = true; //~gfx11! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; f06d0f00 2010540a //~gfx12! image_sample v[84:87], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; e7c6c040 10008054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().a16 = true; //~gfx11! image_sample v[84:85], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D d16 ; f06e0f00 2010540a //~gfx12! image_sample v[84:85], v10, s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_1D d16 ; e7c6c020 10008054 0000000a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1)->mimg().d16 = true; /* NSA */ //~gfx11! image_sample v[84:87], [v10, v40], s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_2D ; f06c0f05 2010540a 00000028 //~gfx12! image_sample v[84:87], [v10, v40], s[64:71], s[32:35] dmask:0xf dim:SQ_RSRC_IMG_2D ; e7c6c001 10008054 0000280a bld.mimg(aco_opcode::image_sample, dst_v4, op_s8, op_s4, Operand(v1), op_v1, Operand(bld.tmp(v1), PhysReg(256 + 40))) ->mimg() .dim = ac_image_2d; //~gfx11! image_bvh_intersect_ray v[84:87], [v40, v42, v[44:46], v[48:50], v[52:54]], s[32:35] ; f0648f81 00085428 34302c2a //~gfx12! image_bvh_intersect_ray v[84:87], [v40, v42, v[44:46], v[48:50], v[52:54]], s[32:35] ; d3c64010 34004054 302c2a28 aco_ptr instr{ create_instruction(aco_opcode::image_bvh_intersect_ray, Format::MIMG, 8, 1)}; instr->definitions[0] = dst_v4; instr->operands[0] = op_s4; instr->operands[1] = Operand(s4); instr->operands[2] = Operand(v1); instr->operands[3] = Operand(PhysReg(256 + 40), v1); /* node */ instr->operands[4] = Operand(PhysReg(256 + 42), v1); /* tmax */ instr->operands[5] = Operand(PhysReg(256 + 44), v3); /* origin */ instr->operands[6] = Operand(PhysReg(256 + 48), v3); /* dir */ instr->operands[7] = Operand(PhysReg(256 + 52), v3); /* inv dir */ instr->mimg().dmask = 0xf; instr->mimg().unrm = true; instr->mimg().r128 = true; bld.insert(std::move(instr)); /* Stores */ //~gfx11! image_store v[30:33], v10, s[64:71] dmask:0xf dim:SQ_RSRC_IMG_1D ; f0180f00 00101e0a //~gfx12! image_store v[30:33], v10, s[64:71] dmask:0xf dim:SQ_RSRC_IMG_1D ; d3c18000 0000801e 0000000a bld.mimg(aco_opcode::image_store, op_s8, Operand(s4), op_v4, op_v1); //~gfx11! image_atomic_add v10, v[20:21], s[64:71] dmask:0x1 dim:SQ_RSRC_IMG_2D ; f0300104 00100a14 //~gfx12! image_atomic_add_uint v10, [v20, v21], s[64:71] dmask:0x1 dim:SQ_RSRC_IMG_2D ; d0430001 0000800a 00001514 bld.mimg(aco_opcode::image_atomic_add, Definition(op_v1.physReg(), v1), op_s8, Operand(s4), op_v1, op_v2, 0x1) ->mimg() .dim = ac_image_2d; /* Atomic with return */ //~gfx11! image_atomic_add v10, v[20:21], s[64:71] dmask:0x1 dim:SQ_RSRC_IMG_2D glc ; f0304104 00100a14 //~gfx12! image_atomic_add_uint v10, [v20, v21], s[64:71] dmask:0x1 dim:SQ_RSRC_IMG_2D th:TH_ATOMIC_RETURN ; d0430001 0010800a 00001514 bld.mimg(aco_opcode::image_atomic_add, Definition(op_v1.physReg(), v1), op_s8, Operand(s4), op_v1, op_v2, 0x1, false, false, false, cache_atomic_rtn) ->mimg() .dim = ac_image_2d; //~gfx11! image_load v[84:87], v[20:21], s[64:71] dmask:0xf dim:SQ_RSRC_IMG_2D ; f0000f04 00105414 //~gfx12! image_load v[84:87], [v20, v21], s[64:71] dmask:0xf dim:SQ_RSRC_IMG_2D ; d3c00001 00008054 00001514 bld.mimg(aco_opcode::image_load, dst_v4, op_s8, Operand(s4), Operand(v1), op_v2)->mimg().dim = ac_image_2d; //~gfx11! image_msaa_load v[84:87], v[30:33], s[64:71] dmask:0x1 dim:SQ_RSRC_IMG_2D_MSAA_ARRAY ; f060011c 0010541e //~gfx12! image_msaa_load v[84:87], [v30, v31, v32, v33], s[64:71] dmask:0x1 dim:SQ_RSRC_IMG_2D_MSAA_ARRAY ; e4460007 00008054 21201f1e bld.mimg(aco_opcode::image_msaa_load, dst_v4, op_s8, Operand(s4), Operand(v1), op_v4, 0x1) ->mimg() .dim = ac_image_2darraymsaa; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.flat) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst_v1 = bld.def(v1); dst_v1.setFixed(PhysReg(256 + 42)); Operand op_s1(bld.tmp(s1)); op_s1.setFixed(PhysReg(32)); Operand op_s2(bld.tmp(s2)); op_s2.setFixed(PhysReg(64)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 10)); Operand op_v2(bld.tmp(v2)); op_v2.setFixed(PhysReg(256 + 20)); /* Addressing */ //~gfx11>> flat_load_b32 v42, v[20:21] ; dc500000 2a7c0014 //~gfx12>> flat_load_b32 v42, v[20:21] ; ec05007c 0000002a 00000014 bld.flat(aco_opcode::flat_load_dword, dst_v1, op_v2, Operand(s1)); //~gfx11! global_load_b32 v42, v[20:21], off ; dc520000 2a7c0014 //~gfx12! global_load_b32 v42, v[20:21], off ; ee05007c 0000002a 00000014 bld.global(aco_opcode::global_load_dword, dst_v1, op_v2, Operand(s1)); //~gfx11! global_load_b32 v42, v10, s[64:65] ; dc520000 2a40000a //~gfx12! global_load_b32 v42, v10, s[64:65] ; ee050040 0000002a 0000000a bld.global(aco_opcode::global_load_dword, dst_v1, op_v1, op_s2); //~gfx11! scratch_load_b32 v42, v10, off ; dc510000 2afc000a //~gfx12! scratch_load_b32 v42, v10, off ; ed05007c 0002002a 0000000a bld.scratch(aco_opcode::scratch_load_dword, dst_v1, op_v1, Operand(s1)); //~gfx11! scratch_load_b32 v42, off, s32 ; dc510000 2a200080 //~gfx12! scratch_load_b32 v42, off, s32 ; ed050020 0000002a 00000000 bld.scratch(aco_opcode::scratch_load_dword, dst_v1, Operand(v1), op_s1); //~gfx11! scratch_load_b32 v42, v10, s32 ; dc510000 2aa0000a //~gfx12! scratch_load_b32 v42, v10, s32 ; ed050020 0002002a 0000000a bld.scratch(aco_opcode::scratch_load_dword, dst_v1, op_v1, op_s1); //~gfx11! scratch_load_b32 v42, off, off ; dc510000 2a7c0080 //~gfx12! scratch_load_b32 v42, off, off ; ed05007c 0000002a 00000000 bld.scratch(aco_opcode::scratch_load_dword, dst_v1, Operand(v1), Operand(s1)); //~gfx11! global_load_b32 v42, v[20:21], off offset:-42 ; dc521fd6 2a7c0014 //~gfx12! global_load_b32 v42, v[20:21], off offset:-42 ; ee05007c 0000002a ffffd614 bld.global(aco_opcode::global_load_dword, dst_v1, op_v2, Operand(s1), -42); //~gfx11! global_load_b32 v42, v[20:21], off offset:84 ; dc520054 2a7c0014 //~gfx12! global_load_b32 v42, v[20:21], off offset:84 ; ee05007c 0000002a 00005414 bld.global(aco_opcode::global_load_dword, dst_v1, op_v2, Operand(s1), 84); /* Various flags */ ac_hw_cache_flags cache_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_sys_coherent = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_non_temporal = {{0, 0, 0, 0, 0}}; ac_hw_cache_flags cache_atomic_rtn = {{0, 0, 0, 0, 0}}; if (gfx >= GFX12) { cache_coherent.gfx12.scope = gfx12_scope_device; cache_sys_coherent.gfx12.scope = gfx12_scope_memory; cache_non_temporal.gfx12.temporal_hint = gfx12_load_non_temporal; cache_atomic_rtn.gfx12.temporal_hint = gfx12_atomic_return; } else { cache_coherent.value = ac_glc; cache_sys_coherent.value = ac_slc; cache_non_temporal.value = ac_dlc; cache_atomic_rtn.value = ac_glc; } //~gfx11! flat_load_b32 v42, v[20:21] slc ; dc508000 2a7c0014 //~gfx12! flat_load_b32 v42, v[20:21] scope:SCOPE_SYS ; ec05007c 000c002a 00000014 bld.flat(aco_opcode::flat_load_dword, dst_v1, op_v2, Operand(s1))->flat().cache = cache_sys_coherent; //~gfx11! flat_load_b32 v42, v[20:21] glc ; dc504000 2a7c0014 //~gfx12! flat_load_b32 v42, v[20:21] scope:SCOPE_DEV ; ec05007c 0008002a 00000014 bld.flat(aco_opcode::flat_load_dword, dst_v1, op_v2, Operand(s1))->flat().cache = cache_coherent; //~gfx11! flat_load_b32 v42, v[20:21] dlc ; dc502000 2a7c0014 //~gfx12! flat_load_b32 v42, v[20:21] th:TH_LOAD_NT ; ec05007c 0010002a 00000014 bld.flat(aco_opcode::flat_load_dword, dst_v1, op_v2, Operand(s1))->flat().cache = cache_non_temporal; /* Stores */ //~gfx11! flat_store_b32 v[20:21], v10 ; dc680000 007c0a14 //~gfx12! flat_store_b32 v[20:21], v10 ; ec06807c 05000000 00000014 bld.flat(aco_opcode::flat_store_dword, op_v2, Operand(s1), op_v1); /* Atomic with return */ //~gfx11! global_atomic_add_u32 v42, v[20:21], v10, off glc ; dcd64000 2a7c0a14 //~gfx12! global_atomic_add_u32 v42, v[20:21], v10, off th:TH_ATOMIC_RETURN ; ee0d407c 0510002a 00000014 bld.global(aco_opcode::global_atomic_add, dst_v1, op_v2, Operand(s1), op_v1)->global().cache = cache_atomic_rtn; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.exp) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Operand op[4]; for (unsigned i = 0; i < 4; i++) op[i] = Operand(PhysReg(256 + i), v1); Operand op_m0(bld.tmp(s1)); op_m0.setFixed(m0); //~gfx11>> exp mrt3 v1, v0, v3, v2 ; f800003f 02030001 //~gfx12>> export mrt3 v1, v0, v3, v2 ; f800003f 02030001 bld.exp(aco_opcode::exp, op[1], op[0], op[3], op[2], 0xf, 3); //~gfx11! exp mrt3 v1, off, v0, off ; f8000035 80008001 //~gfx12! export mrt3 v1, off, v0, off ; f8000035 80008001 bld.exp(aco_opcode::exp, op[1], Operand(v1), op[0], Operand(v1), 0x5, 3); //~gfx11! exp mrt3 v1, v0, v3, v2 done ; f800083f 02030001 //~gfx12! export mrt3 v1, v0, v3, v2 done ; f800083f 02030001 bld.exp(aco_opcode::exp, op[1], op[0], op[3], op[2], 0xf, 3, false, true); //~gfx11! exp mrt3 v1, v0, v3, v2 row_en ; f800203f 02030001 //~gfx12! export mrt3 v1, v0, v3, v2 row_en ; f800203f 02030001 bld.exp(aco_opcode::exp, op[1], op[0], op[3], op[2], op_m0, 0xf, 3)->exp().row_en = true; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.vinterp) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst = bld.def(v1); dst.setFixed(PhysReg(256 + 42)); Operand op0(bld.tmp(v1)); op0.setFixed(PhysReg(256 + 10)); Operand op1(bld.tmp(v1)); op1.setFixed(PhysReg(256 + 20)); Operand op2(bld.tmp(v1)); op2.setFixed(PhysReg(256 + 30)); //! llvm_version: #llvm_ver fprintf(output, "llvm_version: %u\n", LLVM_VERSION_MAJOR); //>> v_interp_p10_f32 v42, v10, v20, v30 wait_exp:7 ; cd00072a 047a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_f32_inreg, dst, op0, op1, op2); //! v_interp_p10_f32 v42, v10, v20, v30 wait_exp:6 ; cd00062a 047a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_f32_inreg, dst, op0, op1, op2, 0, 6); //; if llvm_ver >= 18: //; insert_pattern('v_interp_p2_f32 v42, v10, v20, v30 wait_exp:0 ; cd01002a 047a290a') //; else: //; insert_pattern('v_interp_p2_f32 v42, v10, v20, v30 ; cd01002a 047a290a') bld.vinterp_inreg(aco_opcode::v_interp_p2_f32_inreg, dst, op0, op1, op2, 0, 0); //! v_interp_p10_f32 v42, -v10, v20, v30 wait_exp:6 ; cd00062a 247a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_f32_inreg, dst, op0, op1, op2, 0, 6) ->vinterp_inreg() .neg[0] = true; //! v_interp_p10_f32 v42, v10, -v20, v30 wait_exp:6 ; cd00062a 447a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_f32_inreg, dst, op0, op1, op2, 0, 6) ->vinterp_inreg() .neg[1] = true; //! v_interp_p10_f32 v42, v10, v20, -v30 wait_exp:6 ; cd00062a 847a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_f32_inreg, dst, op0, op1, op2, 0, 6) ->vinterp_inreg() .neg[2] = true; //! v_interp_p10_f16_f32 v42, v10, v20, v30 op_sel:[1,0,0,0] wait_exp:6 ; cd020e2a 047a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_f16_f32_inreg, dst, op0, op1, op2, 0x1, 6); //! v_interp_p2_f16_f32 v42, v10, v20, v30 op_sel:[0,1,0,0] wait_exp:6 ; cd03162a 047a290a bld.vinterp_inreg(aco_opcode::v_interp_p2_f16_f32_inreg, dst, op0, op1, op2, 0x2, 6); //! v_interp_p10_rtz_f16_f32 v42, v10, v20, v30 op_sel:[0,0,1,0] wait_exp:6 ; cd04262a 047a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_rtz_f16_f32_inreg, dst, op0, op1, op2, 0x4, 6); //! v_interp_p2_rtz_f16_f32 v42, v10, v20, v30 op_sel:[0,0,0,1] wait_exp:6 ; cd05462a 047a290a bld.vinterp_inreg(aco_opcode::v_interp_p2_rtz_f16_f32_inreg, dst, op0, op1, op2, 0x8, 6); //! v_interp_p10_f32 v42, v10, v20, v30 clamp wait_exp:6 ; cd00862a 047a290a bld.vinterp_inreg(aco_opcode::v_interp_p10_f32_inreg, dst, op0, op1, op2, 0, 6) ->vinterp_inreg() .clamp = true; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.ldsdir) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst = bld.def(v1); dst.setFixed(PhysReg(256 + 42)); Operand op(bld.tmp(s1)); op.setFixed(m0); //! llvm_version: #llvm_ver fprintf(output, "llvm_version: %u\n", LLVM_VERSION_MAJOR); //~gfx11>> lds_direct_load v42 wait_vdst:15 ; ce1f002a //~gfx12>> ds_direct_load v42 wait_va_vdst:15 wait_vm_vsrc:1 ; ce9f002a bld.ldsdir(aco_opcode::lds_direct_load, dst, op)->ldsdir().wait_vdst = 15; //~gfx11! lds_direct_load v42 wait_vdst:6 ; ce16002a //~gfx12! ds_direct_load v42 wait_va_vdst:6 wait_vm_vsrc:1 ; ce96002a bld.ldsdir(aco_opcode::lds_direct_load, dst, op)->ldsdir().wait_vdst = 6; //; if llvm_ver >= 18 and variant == 'gfx11': //; insert_pattern('lds_direct_load v42 wait_vdst:0 ; ce10002a') //; elif variant == 'gfx11': //; insert_pattern('lds_direct_load v42 ; ce10002a') //~gfx12! ds_direct_load v42 wait_va_vdst:0 wait_vm_vsrc:1 ; ce90002a bld.ldsdir(aco_opcode::lds_direct_load, dst, op)->ldsdir().wait_vdst = 0; //~gfx11! lds_param_load v42, attr56.x wait_vdst:8 ; ce08e02a //~gfx12! ds_param_load v42, attr56.x wait_va_vdst:8 wait_vm_vsrc:1 ; ce88e02a bld.ldsdir(aco_opcode::lds_param_load, dst, op, 56, 0)->ldsdir().wait_vdst = 8; //; if llvm_ver >= 18 and variant == 'gfx11': //; insert_pattern('lds_param_load v42, attr56.x wait_vdst:0 ; ce00e02a') //; elif variant == 'gfx11': //; insert_pattern('lds_param_load v42, attr56.x ; ce00e02a') //~gfx12! ds_param_load v42, attr56.x wait_va_vdst:0 wait_vm_vsrc:1 ; ce80e02a bld.ldsdir(aco_opcode::lds_param_load, dst, op, 56, 0)->ldsdir().wait_vdst = 0; //~gfx11! lds_param_load v42, attr34.y wait_vdst:8 ; ce08892a //~gfx12! ds_param_load v42, attr34.y wait_va_vdst:8 wait_vm_vsrc:1 ; ce88892a bld.ldsdir(aco_opcode::lds_param_load, dst, op, 34, 1)->ldsdir().wait_vdst = 8; //~gfx11! lds_param_load v42, attr12.z wait_vdst:8 ; ce08322a //~gfx12! ds_param_load v42, attr12.z wait_va_vdst:8 wait_vm_vsrc:1 ; ce88322a bld.ldsdir(aco_opcode::lds_param_load, dst, op, 12, 2)->ldsdir().wait_vdst = 8; //~gfx11>> lds_direct_load v42 wait_vdst:15 ; ce1f002a //~gfx12>> ds_direct_load v42 wait_va_vdst:15 wait_vm_vsrc:0 ; ce1f002a bld.ldsdir(aco_opcode::lds_direct_load, dst, op)->ldsdir().wait_vsrc = 0; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.vop12c_v128) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst_v0 = bld.def(v1); dst_v0.setFixed(PhysReg(256)); Definition dst_v128 = bld.def(v1); dst_v128.setFixed(PhysReg(256 + 128)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 1)); Operand op_v2(bld.tmp(v1)); op_v2.setFixed(PhysReg(256 + 2)); Operand op_v129(bld.tmp(v1)); op_v129.setFixed(PhysReg(256 + 129)); Operand op_v130(bld.tmp(v1)); op_v130.setFixed(PhysReg(256 + 130)); //! llvm_version: #llvm_ver fprintf(output, "llvm_version: %u\n", LLVM_VERSION_MAJOR); //>> BB0: //; if llvm_ver == 16: //; insert_pattern('v_mul_f16_e32 v0, v1, v2 ; Error: VGPR_32_Lo128: unknown register 128 ; 6a000501') //; else: //; insert_pattern('v_mul_f16_e32 v0, v1, v2 ; 6a000501') bld.vop2(aco_opcode::v_mul_f16, dst_v0, op_v1, op_v2); //! v_mul_f16_e64 v128, v1, v2 ; d5350080 00020501 bld.vop2(aco_opcode::v_mul_f16, dst_v128, op_v1, op_v2); //! v_mul_f16_e64 v0, v129, v2 ; d5350000 00020581 bld.vop2(aco_opcode::v_mul_f16, dst_v0, op_v129, op_v2); //! v_mul_f16_e64 v0, v1, v130 ; d5350000 00030501 bld.vop2(aco_opcode::v_mul_f16, dst_v0, op_v1, op_v130); //! v_rcp_f16_e64 v128, v1 ; d5d40080 00000101 bld.vop1(aco_opcode::v_rcp_f16, dst_v128, op_v1); //! v_cmp_eq_f16_e64 vcc, v129, v2 ; d402006a 00020581 bld.vopc(aco_opcode::v_cmp_eq_f16, bld.def(s2, vcc), op_v129, op_v2); //! v_mul_f16_e64_dpp v128, v1, v2 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5350080 000204fa ff0d2101 bld.vop2_dpp(aco_opcode::v_mul_f16, dst_v128, op_v1, op_v2, dpp_row_rr(1)); //! v_mul_f16_e64_dpp v0, v129, v2 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5350000 000204fa ff0d2181 bld.vop2_dpp(aco_opcode::v_mul_f16, dst_v0, op_v129, op_v2, dpp_row_rr(1)); //! v_mul_f16_e64_dpp v0, v1, v130 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5350000 000304fa ff0d2101 bld.vop2_dpp(aco_opcode::v_mul_f16, dst_v0, op_v1, op_v130, dpp_row_rr(1)); //! v_mul_f16_e64_dpp v128, v1, v2 dpp8:[0,0,0,0,0,0,0,0] fi:1 ; d5350080 000204ea 00000001 bld.vop2_dpp8(aco_opcode::v_mul_f16, dst_v128, op_v1, op_v2); //! v_mul_f16_e64_dpp v0, v129, v2 dpp8:[0,0,0,0,0,0,0,0] fi:1 ; d5350000 000204ea 00000081 bld.vop2_dpp8(aco_opcode::v_mul_f16, dst_v0, op_v129, op_v2); //! v_mul_f16_e64_dpp v0, v1, v130 dpp8:[0,0,0,0,0,0,0,0] fi:1 ; d5350000 000304ea 00000001 bld.vop2_dpp8(aco_opcode::v_mul_f16, dst_v0, op_v1, op_v130); //! v_fma_f16 v128, v1, v2, 0x60 ; d6480080 03fe0501 00000060 bld.vop2(aco_opcode::v_fmaak_f16, dst_v128, op_v1, op_v2, Operand::literal32(96)); //! v_fma_f16 v128, v1, 0x60, v2 ; d6480080 0409ff01 00000060 bld.vop2(aco_opcode::v_fmamk_f16, dst_v128, op_v1, op_v2, Operand::literal32(96)); //! v_rcp_f16_e64_dpp v128, -v1 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5d40080 200000fa ff1d2101 bld.vop1_dpp(aco_opcode::v_rcp_f16, dst_v128, op_v1, dpp_row_rr(1))->dpp16().neg[0] = true; //! v_rcp_f16_e64_dpp v128, |v1| row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5d40180 000000fa ff2d2101 bld.vop1_dpp(aco_opcode::v_rcp_f16, dst_v128, op_v1, dpp_row_rr(1))->dpp16().abs[0] = true; //! v_mul_f16_e64_dpp v128, -v1, v2 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5350080 200204fa ff1d2101 bld.vop2_dpp(aco_opcode::v_mul_f16, dst_v128, op_v1, op_v2, dpp_row_rr(1))->dpp16().neg[0] = true; //! v_mul_f16_e64_dpp v128, |v1|, v2 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5350180 000204fa ff2d2101 bld.vop2_dpp(aco_opcode::v_mul_f16, dst_v128, op_v1, op_v2, dpp_row_rr(1))->dpp16().abs[0] = true; //! v_cmp_eq_f16_e64_dpp vcc, -v129, v2 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d402006a 200204fa ff1d2181 bld.vopc_dpp(aco_opcode::v_cmp_eq_f16, bld.def(s2, vcc), op_v129, op_v2, dpp_row_rr(1)) ->dpp16() .neg[0] = true; //! v_cmp_eq_f16_e64_dpp vcc, |v129|, v2 row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d402016a 000204fa ff2d2181 bld.vopc_dpp(aco_opcode::v_cmp_eq_f16, bld.def(s2, vcc), op_v129, op_v2, dpp_row_rr(1)) ->dpp16() .abs[0] = true; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.vop3_dpp) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; Definition dst_v0 = bld.def(v1); dst_v0.setFixed(PhysReg(256)); Definition dst_non_vcc = bld.def(s2); dst_non_vcc.setFixed(PhysReg(4)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 1)); Operand op_v2(bld.tmp(v1)); op_v2.setFixed(PhysReg(256 + 2)); Operand op_s1(bld.tmp(s1)); op_s1.setFixed(PhysReg(1)); //>> BB0: //! v_fma_f32_e64_dpp v0, v1, v2, s1 clamp row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d6138000 000604fa ff0d2101 bld.vop3_dpp(aco_opcode::v_fma_f32, dst_v0, op_v1, op_v2, op_s1, dpp_row_rr(1))->valu().clamp = true; //! v_fma_mix_f32_e64_dpp v0, |v1|, |v2|, |s1| op_sel:[1,0,0] op_sel_hi:[1,0,1] row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; cc204f00 080604fa ffad2101 bld.vop3p_dpp(aco_opcode::v_fma_mix_f32, dst_v0, op_v1, op_v2, op_s1, 0x1, 0x5, dpp_row_rr(1)) ->valu() .abs = 0x7; //! v_fma_f32_e64_dpp v0, -v1, -v2, -s1 dpp8:[0,0,0,0,0,0,0,0] fi:1 ; d6130000 e00604ea 00000001 bld.vop3_dpp8(aco_opcode::v_fma_f32, dst_v0, op_v1, op_v2, op_s1)->valu().neg = 0x7; //! v_fma_mix_f32_e64_dpp v0, -v1, -v2, s1 op_sel_hi:[1,1,1] dpp8:[0,0,0,0,0,0,0,0] fi:1 ; cc204000 780604ea 00000001 bld.vop3p_dpp8(aco_opcode::v_fma_mix_f32, dst_v0, op_v1, op_v2, op_s1, 0x0, 0x7)->valu().neg = 0x3; //! v_add_f32_e64_dpp v0, v1, v2 clamp row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5038000 000204fa ff0d2101 bld.vop2_e64_dpp(aco_opcode::v_add_f32, dst_v0, op_v1, op_v2, dpp_row_rr(1))->valu().clamp = true; //! v_sqrt_f32_e64_dpp v0, v1 clamp row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d5b38000 000000fa ff0d2101 bld.vop1_e64_dpp(aco_opcode::v_sqrt_f32, dst_v0, op_v1, dpp_row_rr(1))->valu().clamp = true; //! v_cmp_lt_f32_e64_dpp s[4:5], |v1|, |v2| row_ror:1 row_mask:0xf bank_mask:0xf bound_ctrl:1 fi:1 ; d4110304 000204fa ffad2101 bld.vopc_e64_dpp(aco_opcode::v_cmp_lt_f32, dst_non_vcc, op_v1, op_v2, dpp_row_rr(1)) ->valu() .abs = 0x3; //! v_add_f32_e64_dpp v0, v1, v2 mul:4 dpp8:[0,0,0,0,0,0,0,0] fi:1 ; d5030000 100204ea 00000001 bld.vop2_e64_dpp8(aco_opcode::v_add_f32, dst_v0, op_v1, op_v2)->valu().omod = 2; //! v_sqrt_f32_e64_dpp v0, v1 clamp dpp8:[0,0,0,0,0,0,0,0] fi:1 ; d5b38000 000000ea 00000001 bld.vop1_e64_dpp8(aco_opcode::v_sqrt_f32, dst_v0, op_v1)->valu().clamp = true; //! v_cmp_lt_f32_e64_dpp s[4:5], |v1|, v2 dpp8:[0,0,0,0,0,0,0,0] fi:1 ; d4110104 000204ea 00000001 bld.vopc_e64_dpp8(aco_opcode::v_cmp_lt_f32, dst_non_vcc, op_v1, op_v2)->valu().abs = 0x1; finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.vopd) for (amd_gfx_level gfx : filter_gfx_levels({GFX11, GFX12})) { if (!setup_cs(NULL, gfx)) continue; program->wave_size = 32; Definition dst_v0 = bld.def(v1); dst_v0.setFixed(PhysReg(256)); Definition dst_v1 = bld.def(v1); dst_v1.setFixed(PhysReg(256 + 1)); Operand op_v0(bld.tmp(v1)); op_v0.setFixed(PhysReg(256 + 0)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 1)); Operand op_v2(bld.tmp(v1)); op_v2.setFixed(PhysReg(256 + 2)); Operand op_v3(bld.tmp(v1)); op_v3.setFixed(PhysReg(256 + 3)); Operand op_s0(bld.tmp(s1)); op_s0.setFixed(PhysReg(0)); Operand op_vcc(bld.tmp(s1)); op_vcc.setFixed(vcc); //>> BB0: //! v_dual_mov_b32 v0, v0 :: v_dual_mov_b32 v1, v1 ; ca100100 00000101 bld.vopd(aco_opcode::v_dual_mov_b32, dst_v0, dst_v1, op_v0, op_v1, aco_opcode::v_dual_mov_b32); //! v_dual_mov_b32 v0, 0x60 :: v_dual_mov_b32 v1, s0 ; ca1000ff 00000000 00000060 bld.vopd(aco_opcode::v_dual_mov_b32, dst_v0, dst_v1, Operand::c32(96), op_s0, aco_opcode::v_dual_mov_b32); //! v_dual_mov_b32 v0, s0 :: v_dual_mov_b32 v1, 0x60 ; ca100000 000000ff 00000060 bld.vopd(aco_opcode::v_dual_mov_b32, dst_v0, dst_v1, op_s0, Operand::c32(96), aco_opcode::v_dual_mov_b32); //! v_dual_mul_f32 v0, v0, v1 :: v_dual_mov_b32 v1, v2 ; c8d00300 00000102 bld.vopd(aco_opcode::v_dual_mul_f32, dst_v0, dst_v1, op_v0, op_v1, op_v2, aco_opcode::v_dual_mov_b32); //! v_dual_fmac_f32 v0, v1, v2 :: v_dual_mov_b32 v1, v3 ; c8100501 00000103 bld.vopd(aco_opcode::v_dual_fmac_f32, dst_v0, dst_v1, op_v1, op_v2, op_v0, op_v3, aco_opcode::v_dual_mov_b32); //! v_dual_mov_b32 v0, v0 :: v_dual_and_b32 v1, v1, v2 ; ca240100 00000501 bld.vopd(aco_opcode::v_dual_mov_b32, dst_v0, dst_v1, op_v0, op_v1, op_v2, aco_opcode::v_dual_and_b32); //! v_dual_cndmask_b32 v0, v0, v1 :: v_dual_cndmask_b32 v1, v2, v3 ; ca520300 00000702 bld.vopd(aco_opcode::v_dual_cndmask_b32, dst_v0, dst_v1, op_v0, op_v1, op_vcc, op_v2, op_v3, op_vcc, aco_opcode::v_dual_cndmask_b32); finish_assembler_test(); } END_TEST BEGIN_TEST(assembler.pseudo_scalar_trans) if (LLVM_VERSION_MAJOR < 19 || !setup_cs(NULL, GFX12)) return; //>> v_s_sqrt_f32 s5, s1 ; d6880005 00000001 bld.vop3(aco_opcode::v_s_sqrt_f32, Definition(PhysReg(5), s1), Operand(PhysReg(1), s1)); finish_assembler_test(); END_TEST BEGIN_TEST(assembler.vintrp_high_16bits) for (unsigned i = GFX8; i <= GFX10; i++) { if (!setup_cs(NULL, (amd_gfx_level)i)) continue; Definition dst_v0 = bld.def(v1); dst_v0.setFixed(PhysReg(256)); Definition dst_v1 = bld.def(v1); dst_v1.setFixed(PhysReg(256 + 1)); Operand op_v0(bld.tmp(v1)); op_v0.setFixed(PhysReg(256 + 0)); Operand op_v1(bld.tmp(v1)); op_v1.setFixed(PhysReg(256 + 1)); Operand op_v2(bld.tmp(v1)); op_v2.setFixed(PhysReg(256 + 2)); Operand op_m0(bld.tmp(s1)); op_m0.setFixed(m0); aco_opcode interp_p2_op = aco_opcode::v_interp_p2_f16; if (bld.program->gfx_level == GFX8) interp_p2_op = aco_opcode::v_interp_p2_legacy_f16; //! BB0: //~gfx8! v_interp_p1ll_f16 v0, v1, attr4.y high ; d2740000 00020344 //~gfx9! v_interp_p1ll_f16 v0, v1, attr4.y high ; d2740000 00020344 //~gfx10! v_interp_p1ll_f16 v0, v1, attr4.y high ; d7420000 00020344 bld.vintrp(aco_opcode::v_interp_p1ll_f16, dst_v0, op_v1, op_m0, 4, 1, true); //~gfx8! v_interp_p2_f16 v1, v2, attr4.y, v0 high ; d2760001 04020544 //~gfx9! v_interp_p2_f16 v1, v2, attr4.y, v0 high ; d2770001 04020544 //~gfx10! v_interp_p2_f16 v1, v2, attr4.y, v0 high ; d75a0001 04020544 bld.vintrp(interp_p2_op, dst_v1, op_v2, op_m0, op_v0, 4, 1, true); finish_assembler_test(); } END_TEST