/external/llvm-project/llvm/test/CodeGen/AMDGPU/ |
D | atomic_optimizations_global_pointer.ll | 13 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 15 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 19 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 31 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 33 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 37 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 71 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 73 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 77 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 90 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec [all …]
|
D | fold-reload-into-exec.mir | 39 ; CHECK: S_NOP 0, implicit-def $exec_hi 40 ; CHECK: $sgpr0 = S_MOV_B32 $exec_hi 45 ; CHECK: $exec_hi = S_MOV_B32 killed $sgpr0 47 S_NOP 0, implicit-def $exec_hi 48 %0:sreg_32 = COPY $exec_hi 50 $exec_hi = COPY %0 115 …: S_NOP 0, implicit-def renamable $sgpr0, implicit-def dead renamable $sgpr1, implicit-def $exec_hi 120 ; CHECK: $exec_hi = S_MOV_B32 killed $sgpr0 122 S_NOP 0, implicit-def %0:sreg_32, implicit-def %1:sreg_32, implicit-def $exec_hi 124 $exec_hi = COPY %0
|
D | atomic_optimizations_raw_buffer.ll | 15 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 17 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 21 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 33 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 35 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 39 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 86 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 88 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 92 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 104 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec [all …]
|
D | read_register.ll | 19 ; CHECK: v_mov_b32_e32 v[[HI:[0-9]+]], exec_hi 65 ; CHECK: v_mov_b32_e32 [[COPY:v[0-9]+]], exec_hi 68 %exec_hi = call i32 @llvm.read_register.i32(metadata !6) 69 store i32 %exec_hi, i32 addrspace(1)* %out 81 !6 = !{!"exec_hi"}
|
D | atomic_optimizations_struct_buffer.ll | 15 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 17 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 21 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 33 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 35 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 39 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 99 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 101 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 105 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 117 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec [all …]
|
D | atomic_optimizations_buffer.ll | 16 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 18 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 22 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 34 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 36 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 40 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 118 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec 120 ; GCN64: v_mbcnt_hi_u32_b32{{(_e[0-9]+)?}} v[[mbcnt]], s[[exec_hi]], v[[mbcnt]] 124 ; GCN64: s_bcnt1_i32_b64 s[[popcount:[0-9]+]], s{{\[}}[[exec_lo]]:[[exec_hi]]{{\]}} 136 ; GCN64: s_mov_b64 s{{\[}}[[exec_lo:[0-9]+]]:[[exec_hi:[0-9]+]]{{\]}}, exec [all …]
|
D | write_register.ll | 82 ; CHECK: s_mov_b32 exec_hi, 0 83 ; CHECK: s_mov_b32 exec_hi, s{{[0-9]+}} 102 !6 = !{!"exec_hi"}
|
D | llvm.amdgcn.ballot.i64.ll | 24 ; CHECK-NEXT: s_mov_b32 s1, exec_hi
|
D | atomic_optimizations_local_pointer.ll | 376 ; GFX8-NEXT: v_mbcnt_hi_u32_b32 v0, exec_hi, v0 427 ; GFX9-NEXT: v_mbcnt_hi_u32_b32 v0, exec_hi, v0 501 ; GFX1064-NEXT: v_mbcnt_hi_u32_b32_e64 v0, exec_hi, v0 615 ; GFX8-NEXT: v_mbcnt_hi_u32_b32 v0, exec_hi, v0 666 ; GFX9-NEXT: v_mbcnt_hi_u32_b32 v0, exec_hi, v0 740 ; GFX1064-NEXT: v_mbcnt_hi_u32_b32_e64 v0, exec_hi, v0 854 ; GFX8-NEXT: v_mbcnt_hi_u32_b32 v0, exec_hi, v0 905 ; GFX9-NEXT: v_mbcnt_hi_u32_b32 v0, exec_hi, v0 979 ; GFX1064-NEXT: v_mbcnt_hi_u32_b32_e64 v0, exec_hi, v0 1920 ; GFX8-NEXT: v_mbcnt_hi_u32_b32 v0, exec_hi, v0 [all …]
|
/external/llvm-project/llvm/test/MC/AMDGPU/ |
D | smem-err.s | 15 s_store_dword exec_hi, s[2:3], 0x0 27 s_buffer_store_dword exec_hi, s[0:3], 0x0 39 s_load_dword exec_hi, s[0:1], s4 51 s_buffer_load_dword exec_hi, s[0:3], s4
|
D | gfx9_unsupported.s | 61 s_and_saveexec_b32 exec_hi, s1 64 s_andn1_saveexec_b32 exec_hi, s1 67 s_andn1_wrexec_b32 exec_hi, s1 70 s_andn2_saveexec_b32 exec_hi, s1 73 s_andn2_wrexec_b32 exec_hi, s1 97 s_nand_saveexec_b32 exec_hi, s1 100 s_nor_saveexec_b32 exec_hi, s1 103 s_or_saveexec_b32 exec_hi, s1 106 s_orn1_saveexec_b32 exec_hi, s1 109 s_orn2_saveexec_b32 exec_hi, s1 [all …]
|
D | gfx8_unsupported.s | 289 s_and_saveexec_b32 exec_hi, s1 292 s_andn1_saveexec_b32 exec_hi, s1 298 s_andn1_wrexec_b32 exec_hi, s1 304 s_andn2_saveexec_b32 exec_hi, s1 307 s_andn2_wrexec_b32 exec_hi, s1 502 s_lshl1_add_u32 exec_hi, s1, s2 505 s_lshl2_add_u32 exec_hi, s1, s2 508 s_lshl3_add_u32 exec_hi, s1, s2 511 s_lshl4_add_u32 exec_hi, s1, s2 517 s_mul_hi_i32 exec_hi, s1, s2 [all …]
|
D | gfx7_unsupported.s | 295 s_and_saveexec_b32 exec_hi, s1 298 s_andn1_saveexec_b32 exec_hi, s1 304 s_andn1_wrexec_b32 exec_hi, s1 310 s_andn2_saveexec_b32 exec_hi, s1 313 s_andn2_wrexec_b32 exec_hi, s1 484 s_buffer_store_dword exec_hi, s[0:3], 0x0 538 s_lshl1_add_u32 exec_hi, s1, s2 541 s_lshl2_add_u32 exec_hi, s1, s2 544 s_lshl3_add_u32 exec_hi, s1, s2 547 s_lshl4_add_u32 exec_hi, s1, s2 [all …]
|
D | vop-err.s | 27 v_movreld_b32 v0, exec_hi 108 v_cndmask_b32 v0, exec_hi, v2, vcc 168 v_addc_u32 v0, vcc, exec_hi, v0, vcc 220 v_madak_f32 v0, exec_hi, v0, 0x11213141
|
D | flat-gfx10.s | 91 flat_load_dword v1, v[3:4], exec_hi 94 flat_store_dword v[3:4], v1, exec_hi
|
D | flat-gfx9.s | 114 flat_load_dword v1, v[3:4], exec_hi 117 flat_store_dword v[3:4], v1, exec_hi
|
D | flat-scratch-instructions.s | 244 scratch_load_dword v1, off, exec_hi 249 scratch_store_dword off, v2, exec_hi
|
D | reg-syntax-extra.s | 35 s_mov_b64 [exec_lo,exec_hi], s[2:3]
|
D | dl-insts.s | 31 v_fmac_f32 v5, exec_hi, v2 70 v_fmac_f32_e64 v5, exec_hi, v2 98 v_fmac_f32_e64 v5, v1, exec_hi 216 v_xnor_b32 v5, exec_hi, v2 255 v_xnor_b32_e64 v5, exec_hi, v2 283 v_xnor_b32_e64 v5, v1, exec_hi 316 v_xnor_b32_sdwa v5, exec_hi, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
|
D | reg-syntax-err.s | 78 s_mov_b64 s[10:11], [exec_hi,exec_lo]
|
/external/llvm/test/CodeGen/AMDGPU/ |
D | read_register.ll | 19 ; CHECK: v_mov_b32_e32 v[[HI:[0-9]+]], exec_hi 65 ; CHECK: v_mov_b32_e32 [[COPY:v[0-9]+]], exec_hi 68 %exec_hi = call i32 @llvm.read_register.i32(metadata !6) 69 store i32 %exec_hi, i32 addrspace(1)* %out 81 !6 = !{!"exec_hi"}
|
D | write_register.ll | 64 ; CHECK: s_mov_b32 exec_hi, 0 65 ; CHECK: s_mov_b32 exec_hi, s{{[0-9]+}} 80 !6 = !{!"exec_hi"}
|
/external/llvm/test/MC/AMDGPU/ |
D | reg-syntax-extra.s | 26 s_mov_b64 [exec_lo,exec_hi], s[2:3]
|
/external/llvm-project/llvm/test/MC/Disassembler/AMDGPU/ |
D | gfx10_dasm_all.txt | 8366 # GFX10: s_abs_i32 exec_hi, s1 ; encoding: [0x01,0x34,0xff,0xbe] 8393 # GFX10: s_abs_i32 s0, exec_hi ; encoding: [0x7f,0x34,0x80,0xbe] 8426 # GFX10: s_absdiff_i32 exec_hi, s1, s2 ; encoding: [0x01,0x02,0x7f,0x96] 8453 # GFX10: s_absdiff_i32 s0, exec_hi, s2 ; encoding: [0x7f,0x02,0x00,0x96] 8480 # GFX10: s_absdiff_i32 s0, s1, exec_hi ; encoding: [0x01,0x7f,0x00,0x96] 8531 # GFX10: s_add_i32 exec_hi, s1, s2 ; encoding: [0x01,0x02,0x7f,0x81] 8558 # GFX10: s_add_i32 s0, exec_hi, s2 ; encoding: [0x7f,0x02,0x00,0x81] 8585 # GFX10: s_add_i32 s0, s1, exec_hi ; encoding: [0x01,0x7f,0x00,0x81] 8636 # GFX10: s_add_u32 exec_hi, s1, s2 ; encoding: [0x01,0x02,0x7f,0x80] 8663 # GFX10: s_add_u32 s0, exec_hi, s2 ; encoding: [0x7f,0x02,0x00,0x80] [all …]
|
/external/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/ |
D | llvm.amdgcn.ballot.i64.ll | 24 ; CHECK-NEXT: s_mov_b32 s1, exec_hi
|