; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1031 -verify-coalescing < %s | FileCheck %s
define amdgpu_kernel void @blender_no_live_segment_at_def_error(<4 x float> %extractVec358.i.i, i32 %cmp5.i.i.arg, float %i1.i, i32 %cmp221.i.i.arg, i32 %cmp262.i.i.arg, ptr addrspace(1) %arg) {
; CHECK-LABEL: blender_no_live_segment_at_def_error:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: s_add_u32 s10, s10, s15
; CHECK-NEXT: s_addc_u32 s11, s11, 0
; CHECK-NEXT: s_mov_b32 s32, 0
; CHECK-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s10
; CHECK-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s11
; CHECK-NEXT: s_load_dwordx8 s[36:43], s[6:7], 0x0
; CHECK-NEXT: s_add_u32 s0, s0, s15
; CHECK-NEXT: s_addc_u32 s1, s1, 0
; CHECK-NEXT: s_mov_b64 s[10:11], s[8:9]
; CHECK-NEXT: s_mov_b32 s8, 0
; CHECK-NEXT: s_waitcnt lgkmcnt(0)
; CHECK-NEXT: s_cmp_lg_u32 s40, 0
; CHECK-NEXT: s_cbranch_scc1 .LBB0_8
; CHECK-NEXT: ; %bb.1: ; %if.end13.i.i
; CHECK-NEXT: s_cmp_eq_u32 s42, 0
; CHECK-NEXT: s_cbranch_scc1 .LBB0_4
; CHECK-NEXT: ; %bb.2: ; %if.else251.i.i
; CHECK-NEXT: s_cmp_lg_u32 s43, 0
; CHECK-NEXT: s_mov_b32 s15, 0
; CHECK-NEXT: s_cselect_b32 s8, -1, 0
; CHECK-NEXT: s_and_b32 vcc_lo, exec_lo, s8
; CHECK-NEXT: s_cbranch_vccz .LBB0_5
; CHECK-NEXT: ; %bb.3:
; CHECK-NEXT: s_mov_b32 s36, 0
; CHECK-NEXT: s_andn2_b32 vcc_lo, exec_lo, s8
; CHECK-NEXT: s_cbranch_vccz .LBB0_6
; CHECK-NEXT: s_branch .LBB0_7
; CHECK-NEXT: .LBB0_4:
; CHECK-NEXT: s_mov_b32 s10, s8
; CHECK-NEXT: s_mov_b32 s11, s8
; CHECK-NEXT: s_mov_b32 s9, s8
; CHECK-NEXT: s_mov_b64 s[38:39], s[10:11]
; CHECK-NEXT: s_mov_b64 s[36:37], s[8:9]
; CHECK-NEXT: s_branch .LBB0_7
; CHECK-NEXT: .LBB0_5: ; %if.then263.i.i
; CHECK-NEXT: v_cmp_lt_f32_e64 s8, s41, 0
; CHECK-NEXT: s_mov_b32 s36, 1.0
; CHECK-NEXT: s_mov_b32 s15, 0x7fc00000
; CHECK-NEXT: s_mov_b32 s37, s36
; CHECK-NEXT: s_mov_b32 s38, s36
; CHECK-NEXT: s_mov_b32 s39, s36
; CHECK-NEXT: s_andn2_b32 vcc_lo, exec_lo, s8
; CHECK-NEXT: s_cbranch_vccnz .LBB0_7
; CHECK-NEXT: .LBB0_6: ; %if.end273.i.i
; CHECK-NEXT: s_add_u32 s8, s6, 40
; CHECK-NEXT: s_addc_u32 s9, s7, 0
; CHECK-NEXT: s_getpc_b64 s[16:17]
; CHECK-NEXT: s_add_u32 s16, s16, _Z3dotDv3_fS_@gotpcrel32@lo+4
; CHECK-NEXT: s_addc_u32 s17, s17, _Z3dotDv3_fS_@gotpcrel32@hi+12
; CHECK-NEXT: v_lshlrev_b32_e32 v2, 20, v2
; CHECK-NEXT: s_load_dwordx2 s[16:17], s[16:17], 0x0
; CHECK-NEXT: v_lshlrev_b32_e32 v3, 10, v1
; CHECK-NEXT: v_add_f32_e64 v1, s15, s36
; CHECK-NEXT: s_mov_b32 s36, 0
; CHECK-NEXT: s_mov_b64 s[34:35], s[6:7]
; CHECK-NEXT: v_or3_b32 v31, v0, v3, v2
; CHECK-NEXT: v_mov_b32_e32 v0, v1
; CHECK-NEXT: v_mov_b32_e32 v1, 0
; CHECK-NEXT: v_mov_b32_e32 v2, 0
; CHECK-NEXT: s_waitcnt lgkmcnt(0)
; CHECK-NEXT: s_swappc_b64 s[30:31], s[16:17]
; CHECK-NEXT: s_mov_b64 s[6:7], s[34:35]
; CHECK-NEXT: s_mov_b32 s37, s36
; CHECK-NEXT: s_mov_b32 s38, s36
; CHECK-NEXT: s_mov_b32 s39, s36
; CHECK-NEXT: .LBB0_7: ; %if.end294.i.i
; CHECK-NEXT: v_mov_b32_e32 v0, 0
; CHECK-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:12
; CHECK-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:8
; CHECK-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:4
; CHECK-NEXT: buffer_store_dword v0, off, s[0:3], 0
; CHECK-NEXT: .LBB0_8: ; %kernel_direct_lighting.exit
; CHECK-NEXT: s_load_dwordx2 s[4:5], s[6:7], 0x20
; CHECK-NEXT: v_mov_b32_e32 v0, s36
; CHECK-NEXT: v_mov_b32_e32 v4, 0
; CHECK-NEXT: v_mov_b32_e32 v1, s37
; CHECK-NEXT: v_mov_b32_e32 v2, s38
; CHECK-NEXT: v_mov_b32_e32 v3, s39
; CHECK-NEXT: s_waitcnt lgkmcnt(0)
; CHECK-NEXT: global_store_dwordx4 v4, v[0:3], s[4:5]
; CHECK-NEXT: s_endpgm
entry:
%cmp5.i.i = icmp eq i32 %cmp5.i.i.arg, 0
br i1 %cmp5.i.i, label %if.end13.i.i, label %kernel_direct_lighting.exit
if.end13.i.i: ; preds = %entry
%cmp221.i.i = icmp eq i32 %cmp221.i.i.arg, 0
br i1 %cmp221.i.i, label %if.end294.i.i, label %if.else251.i.i
if.else251.i.i: ; preds = %if.end13.i.i
%cmp262.i.i = icmp eq i32 %cmp262.i.i.arg, 0
br i1 %cmp262.i.i, label %if.then263.i.i, label %if.end273.i.i
if.then263.i.i: ; preds = %if.else251.i.i
%cmp.i604.i.i = fcmp olt float %i1.i, 0.000000e+00
br i1 %cmp.i604.i.i, label %if.end273.i.i, label %if.end294.i.i
if.end273.i.i: ; preds = %if.then263.i.i, %if.else251.i.i
%i = phi float [ 1.000000e+00, %if.then263.i.i ], [ 0.000000e+00, %if.else251.i.i ]
%i1 = phi float [ 0x7FF8000000000000, %if.then263.i.i ], [ 0.000000e+00, %if.else251.i.i ]
%extractVec278.i.i = insertelement <3 x float> zeroinitializer, float %i1, i64 0
%extractVec3.i.i.i = insertelement <3 x float> zeroinitializer, float %i, i64 0
%i3.i = fadd <3 x float> %extractVec278.i.i, %extractVec3.i.i.i
%call32.i.i.i = tail call float @_Z3dotDv3_fS_(<3 x float> %i3.i)
br label %if.end294.i.i
if.end294.i.i: ; preds = %if.end273.i.i, %if.then263.i.i, %if.end13.i.i
%ls111.sroa.0.2.i = phi <4 x float> [ zeroinitializer, %if.end13.i.i ], [ zeroinitializer, %if.end273.i.i ], [ <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %if.then263.i.i ]
store <4 x float> zeroinitializer, ptr addrspace(5) null, align 4
br label %kernel_direct_lighting.exit
kernel_direct_lighting.exit: ; preds = %if.end294.i.i, %entry
%ls111.sroa.0.3.i = phi <4 x float> [ %ls111.sroa.0.2.i, %if.end294.i.i ], [ %extractVec358.i.i, %entry ]
store <4 x float> %ls111.sroa.0.3.i, ptr addrspace(1) %arg, align 4
ret void
}
declare float @_Z3dotDv3_fS_(<3 x float>)
!llvm.module.flags = !{!0}
!0 = !{i32 1, !"amdhsa_code_object_version", i32 500}
Codebase Browser
llvm