| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5 |
| ; RUN: llc -global-isel=0 -mtriple=amdgpu12.50 < %s | FileCheck -check-prefixes=GFX12,GFX12-SDAG %s |
| ; RUN: llc -global-isel=0 -mtriple=amdgpu12.50 < %s | FileCheck -check-prefixes=GFX12,GFX12-ISEL %s |
| |
| define amdgpu_kernel void @kernel1() #0 { |
| ; GFX12-LABEL: kernel1: |
| ; GFX12: ; %bb.0: |
| ; GFX12-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_WAVE_MODE, 25, 1), 1 ; msbs: dst=0 src0=0 src1=0 src2=0 |
| ; GFX12-NEXT: s_mov_b64 s[64:65], 0 |
| ; GFX12-NEXT: v_nop |
| ; GFX12-NEXT: global_prefetch_b8 v0, s[64:65] scope:SCOPE_SE |
| ; GFX12-NEXT: s_cmp_eq_u32 0, 0 |
| ; GFX12-NEXT: s_barrier_signal_isfirst -1 |
| ; GFX12-NEXT: s_barrier_wait -1 |
| ; GFX12-NEXT: s_cselect_b32 s0, 1, 0 |
| ; GFX12-NEXT: s_delay_alu instid0(SALU_CYCLE_1) |
| ; GFX12-NEXT: s_cmp_lg_u32 s0, 1 |
| ; GFX12-NEXT: s_cbranch_scc1 .LBB0_2 |
| ; GFX12-NEXT: ; %bb.1: |
| ; GFX12-NEXT: s_barrier_signal -3 |
| ; GFX12-NEXT: .LBB0_2: |
| ; GFX12-NEXT: s_barrier_wait -3 |
| ; GFX12-NEXT: s_get_barrier_state s0, -3 |
| ; GFX12-NEXT: s_endpgm |
| call void @llvm.amdgcn.s.cluster.barrier() |
| %state3 = call i32 @llvm.amdgcn.s.get.barrier.state(i32 -3) |
| ret void |
| } |
| |
| declare void @llvm.amdgcn.s.cluster.barrier() #1 |
| declare i32 @llvm.amdgcn.s.get.barrier.state(i32) #1 |
| |
| attributes #0 = { nounwind } |
| attributes #1 = { convergent nounwind } |
| attributes #2 = { nounwind readnone } |
| ;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line: |
| ; GFX12-ISEL: {{.*}} |
| ; GFX12-SDAG: {{.*}} |
