THU-DSP-LAB
/
llvm-project
mirror of https://github.com/THU-DSP-LAB/llvm-project.git
2
0
Fork 1
Code Issues Packages Projects Releases Wiki Activity
llvm-project/llvm/test/CodeGen/AMDGPU/fabs.f16.ll

595 lines
22 KiB
LLVM
Raw Permalink Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn--amdhsa -mcpu=kaveri -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,CI %s
; RUN: llc -mtriple=amdgcn--amdhsa -mcpu=tonga -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX89,VI %s
; RUN: llc -mtriple=amdgcn--amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX89,GFX9 %s
; DAGCombiner will transform:
; (fabs (f16 bitcast (i16 a))) => (f16 bitcast (and (i16 a), 0x7FFFFFFF))
; unless isFabsFree returns true
define amdgpu_kernel void @s_fabs_free_f16(ptr addrspace(1) %out, i16 %in) {
; CI-LABEL: s_fabs_free_f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dword s2, s[4:5], 0x2
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: s_and_b32 s2, s2, 0x7fff
; CI-NEXT: v_mov_b32_e32 v0, s0
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_mov_b32_e32 v2, s2
; CI-NEXT: flat_store_short v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: s_fabs_free_f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dword s2, s[4:5], 0x8
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: s_and_b32 s2, s2, 0x7fff
; VI-NEXT: v_mov_b32_e32 v0, s0
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: v_mov_b32_e32 v2, s2
; VI-NEXT: flat_store_short v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: s_fabs_free_f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s2, s[4:5], 0x8
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_and_b32 s2, s2, 0x7fff
; GFX9-NEXT: v_mov_b32_e32 v1, s2
; GFX9-NEXT: global_store_short v0, v1, s[0:1]
; GFX9-NEXT: s_endpgm
%bc= bitcast i16 %in to half
%fabs = call half @llvm.fabs.f16(half %bc)
store half %fabs, ptr addrspace(1) %out
ret void
}
define amdgpu_kernel void @s_fabs_f16(ptr addrspace(1) %out, half %in) {
; CI-LABEL: s_fabs_f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dword s2, s[4:5], 0x2
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: s_and_b32 s2, s2, 0x7fff
; CI-NEXT: v_mov_b32_e32 v0, s0
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_mov_b32_e32 v2, s2
; CI-NEXT: flat_store_short v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: s_fabs_f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dword s2, s[4:5], 0x8
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: s_and_b32 s2, s2, 0x7fff
; VI-NEXT: v_mov_b32_e32 v0, s0
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: v_mov_b32_e32 v2, s2
; VI-NEXT: flat_store_short v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: s_fabs_f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s2, s[4:5], 0x8
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_and_b32 s2, s2, 0x7fff
; GFX9-NEXT: v_mov_b32_e32 v1, s2
; GFX9-NEXT: global_store_short v0, v1, s[0:1]
; GFX9-NEXT: s_endpgm
%fabs = call half @llvm.fabs.f16(half %in)
store half %fabs, ptr addrspace(1) %out
ret void
}
define amdgpu_kernel void @s_fabs_v2f16(ptr addrspace(1) %out, <2 x half> %in) {
; CI-LABEL: s_fabs_v2f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dword s2, s[4:5], 0x2
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; CI-NEXT: v_mov_b32_e32 v0, s0
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_mov_b32_e32 v2, s2
; CI-NEXT: flat_store_dword v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: s_fabs_v2f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dword s2, s[4:5], 0x8
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; VI-NEXT: v_mov_b32_e32 v0, s0
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: v_mov_b32_e32 v2, s2
; VI-NEXT: flat_store_dword v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: s_fabs_v2f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s2, s[4:5], 0x8
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; GFX9-NEXT: v_mov_b32_e32 v1, s2
; GFX9-NEXT: global_store_dword v0, v1, s[0:1]
; GFX9-NEXT: s_endpgm
%fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %in)
store <2 x half> %fabs, ptr addrspace(1) %out
ret void
}
define amdgpu_kernel void @s_fabs_v4f16(ptr addrspace(1) %out, <4 x half> %in) {
; CI-LABEL: s_fabs_v4f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: s_and_b32 s3, s3, 0x7fff7fff
; CI-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; CI-NEXT: v_mov_b32_e32 v3, s1
; CI-NEXT: v_mov_b32_e32 v0, s2
; CI-NEXT: v_mov_b32_e32 v1, s3
; CI-NEXT: v_mov_b32_e32 v2, s0
; CI-NEXT: flat_store_dwordx2 v[2:3], v[0:1]
; CI-NEXT: s_endpgm
;
; VI-LABEL: s_fabs_v4f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: s_and_b32 s3, s3, 0x7fff7fff
; VI-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; VI-NEXT: v_mov_b32_e32 v3, s1
; VI-NEXT: v_mov_b32_e32 v0, s2
; VI-NEXT: v_mov_b32_e32 v1, s3
; VI-NEXT: v_mov_b32_e32 v2, s0
; VI-NEXT: flat_store_dwordx2 v[2:3], v[0:1]
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: s_fabs_v4f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GFX9-NEXT: v_mov_b32_e32 v2, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_and_b32 s3, s3, 0x7fff7fff
; GFX9-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; GFX9-NEXT: v_mov_b32_e32 v0, s2
; GFX9-NEXT: v_mov_b32_e32 v1, s3
; GFX9-NEXT: global_store_dwordx2 v2, v[0:1], s[0:1]
; GFX9-NEXT: s_endpgm
%fabs = call <4 x half> @llvm.fabs.v4f16(<4 x half> %in)
store <4 x half> %fabs, ptr addrspace(1) %out
ret void
}
define amdgpu_kernel void @fabs_fold_f16(ptr addrspace(1) %out, half %in0, half %in1) {
; CI-LABEL: fabs_fold_f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dword s0, s[4:5], 0x2
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: v_cvt_f32_f16_e64 v0, |s0|
; CI-NEXT: s_lshr_b32 s0, s0, 16
; CI-NEXT: v_cvt_f32_f16_e32 v1, s0
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; CI-NEXT: v_mul_f32_e32 v0, v0, v1
; CI-NEXT: v_cvt_f16_f32_e32 v2, v0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: v_mov_b32_e32 v0, s0
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: flat_store_short v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: fabs_fold_f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dword s2, s[4:5], 0x8
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: s_lshr_b32 s3, s2, 16
; VI-NEXT: v_mov_b32_e32 v0, s3
; VI-NEXT: v_mul_f16_e64 v2, |s2|, v0
; VI-NEXT: v_mov_b32_e32 v0, s0
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: flat_store_short v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: fabs_fold_f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s2, s[4:5], 0x8
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_lshr_b32 s3, s2, 16
; GFX9-NEXT: v_mov_b32_e32 v1, s3
; GFX9-NEXT: v_mul_f16_e64 v1, |s2|, v1
; GFX9-NEXT: global_store_short v0, v1, s[0:1]
; GFX9-NEXT: s_endpgm
%fabs = call half @llvm.fabs.f16(half %in0)
%fmul = fmul half %fabs, %in1
store half %fmul, ptr addrspace(1) %out
ret void
}
define amdgpu_kernel void @v_fabs_v2f16(ptr addrspace(1) %out, ptr addrspace(1) %in) #0 {
; CI-LABEL: v_fabs_v2f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x2
; CI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_add_i32_e32 v0, vcc, s0, v0
; CI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT: flat_load_dword v2, v[0:1]
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: v_and_b32_e32 v2, 0x7fff7fff, v2
; CI-NEXT: flat_store_dword v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: v_fabs_v2f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x8
; VI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: v_add_u32_e32 v0, vcc, s0, v0
; VI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT: flat_load_dword v2, v[0:1]
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: v_and_b32_e32 v2, 0x7fff7fff, v2
; VI-NEXT: flat_store_dword v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: v_fabs_v2f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x8
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: global_load_dword v1, v0, s[0:1]
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_and_b32_e32 v1, 0x7fff7fff, v1
; GFX9-NEXT: global_store_dword v0, v1, s[0:1]
; GFX9-NEXT: s_endpgm
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.in = getelementptr inbounds <2 x half>, ptr addrspace(1) %in, i32 %tid
%gep.out = getelementptr inbounds <2 x half>, ptr addrspace(1) %in, i32 %tid
%val = load <2 x half>, ptr addrspace(1) %gep.in, align 2
%fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %val)
store <2 x half> %fabs, ptr addrspace(1) %gep.out
ret void
}
define amdgpu_kernel void @fabs_free_v2f16(ptr addrspace(1) %out, i32 %in) #0 {
; CI-LABEL: fabs_free_v2f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dword s2, s[4:5], 0x2
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; CI-NEXT: v_mov_b32_e32 v0, s0
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_mov_b32_e32 v2, s2
; CI-NEXT: flat_store_dword v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: fabs_free_v2f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dword s2, s[4:5], 0x8
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; VI-NEXT: v_mov_b32_e32 v0, s0
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: v_mov_b32_e32 v2, s2
; VI-NEXT: flat_store_dword v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: fabs_free_v2f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s2, s[4:5], 0x8
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; GFX9-NEXT: v_mov_b32_e32 v0, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_and_b32 s2, s2, 0x7fff7fff
; GFX9-NEXT: v_mov_b32_e32 v1, s2
; GFX9-NEXT: global_store_dword v0, v1, s[0:1]
; GFX9-NEXT: s_endpgm
%bc = bitcast i32 %in to <2 x half>
%fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %bc)
store <2 x half> %fabs, ptr addrspace(1) %out
ret void
}
; FIXME: Should do fabs after conversion to avoid converting multiple
; times in this particular case.
define amdgpu_kernel void @v_fabs_fold_self_v2f16(ptr addrspace(1) %out, ptr addrspace(1) %in) #0 {
; CI-LABEL: v_fabs_fold_self_v2f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; CI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: v_mov_b32_e32 v1, s3
; CI-NEXT: v_add_i32_e32 v0, vcc, s2, v0
; CI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT: flat_load_dword v0, v[0:1]
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: v_lshrrev_b32_e32 v1, 16, v0
; CI-NEXT: v_cvt_f32_f16_e32 v2, v1
; CI-NEXT: v_cvt_f32_f16_e64 v1, |v1|
; CI-NEXT: v_cvt_f32_f16_e32 v3, v0
; CI-NEXT: v_cvt_f32_f16_e64 v0, |v0|
; CI-NEXT: v_mul_f32_e32 v1, v1, v2
; CI-NEXT: v_cvt_f16_f32_e32 v2, v1
; CI-NEXT: v_mul_f32_e32 v0, v0, v3
; CI-NEXT: v_cvt_f16_f32_e32 v3, v0
; CI-NEXT: v_mov_b32_e32 v0, s0
; CI-NEXT: v_lshlrev_b32_e32 v2, 16, v2
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_or_b32_e32 v2, v3, v2
; CI-NEXT: flat_store_dword v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: v_fabs_fold_self_v2f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; VI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: v_mov_b32_e32 v1, s3
; VI-NEXT: v_add_u32_e32 v0, vcc, s2, v0
; VI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT: flat_load_dword v2, v[0:1]
; VI-NEXT: v_mov_b32_e32 v0, s0
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: v_mul_f16_sdwa v3, |v2|, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; VI-NEXT: v_mul_f16_e64 v2, |v2|, v2
; VI-NEXT: v_or_b32_e32 v2, v2, v3
; VI-NEXT: flat_store_dword v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: v_fabs_fold_self_v2f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: v_mov_b32_e32 v1, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: global_load_dword v0, v0, s[2:3]
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_and_b32_e32 v2, 0x7fff7fff, v0
; GFX9-NEXT: v_pk_mul_f16 v0, v2, v0
; GFX9-NEXT: global_store_dword v1, v0, s[0:1]
; GFX9-NEXT: s_endpgm
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep = getelementptr <2 x half>, ptr addrspace(1) %in, i32 %tid
%val = load <2 x half>, ptr addrspace(1) %gep
%fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %val)
%fmul = fmul <2 x half> %fabs, %val
store <2 x half> %fmul, ptr addrspace(1) %out
ret void
}
define amdgpu_kernel void @v_fabs_fold_v2f16(ptr addrspace(1) %out, ptr addrspace(1) %in, i32 %other.val) #0 {
; CI-LABEL: v_fabs_fold_v2f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; CI-NEXT: s_load_dword s4, s[4:5], 0x4
; CI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: v_mov_b32_e32 v1, s3
; CI-NEXT: v_add_i32_e32 v0, vcc, s2, v0
; CI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT: flat_load_dword v0, v[0:1]
; CI-NEXT: s_lshr_b32 s2, s4, 16
; CI-NEXT: v_cvt_f32_f16_e32 v1, s2
; CI-NEXT: v_cvt_f32_f16_e32 v3, s4
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: v_lshrrev_b32_e32 v2, 16, v0
; CI-NEXT: v_cvt_f32_f16_e64 v2, |v2|
; CI-NEXT: v_cvt_f32_f16_e64 v0, |v0|
; CI-NEXT: v_mul_f32_e32 v1, v2, v1
; CI-NEXT: v_cvt_f16_f32_e32 v2, v1
; CI-NEXT: v_mul_f32_e32 v0, v0, v3
; CI-NEXT: v_cvt_f16_f32_e32 v3, v0
; CI-NEXT: v_mov_b32_e32 v0, s0
; CI-NEXT: v_lshlrev_b32_e32 v2, 16, v2
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_or_b32_e32 v2, v3, v2
; CI-NEXT: flat_store_dword v[0:1], v2
; CI-NEXT: s_endpgm
;
; VI-LABEL: v_fabs_fold_v2f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; VI-NEXT: s_load_dword s4, s[4:5], 0x10
; VI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: v_mov_b32_e32 v1, s3
; VI-NEXT: v_add_u32_e32 v0, vcc, s2, v0
; VI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT: flat_load_dword v2, v[0:1]
; VI-NEXT: v_mov_b32_e32 v0, s0
; VI-NEXT: s_lshr_b32 s0, s4, 16
; VI-NEXT: v_mov_b32_e32 v3, s0
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: v_mul_f16_sdwa v3, |v2|, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; VI-NEXT: v_mul_f16_e64 v2, |v2|, s4
; VI-NEXT: v_or_b32_e32 v2, v2, v3
; VI-NEXT: flat_store_dword v[0:1], v2
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: v_fabs_fold_v2f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GFX9-NEXT: s_load_dword s6, s[4:5], 0x10
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: v_mov_b32_e32 v1, 0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: global_load_dword v0, v0, s[2:3]
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_and_b32_e32 v0, 0x7fff7fff, v0
; GFX9-NEXT: v_pk_mul_f16 v0, v0, s6
; GFX9-NEXT: global_store_dword v1, v0, s[0:1]
; GFX9-NEXT: s_endpgm
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep = getelementptr <2 x half>, ptr addrspace(1) %in, i32 %tid
%val = load <2 x half>, ptr addrspace(1) %gep
%fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %val)
%other.val.cvt = bitcast i32 %other.val to <2 x half>
%fmul = fmul <2 x half> %fabs, %other.val.cvt
store <2 x half> %fmul, ptr addrspace(1) %out
ret void
}
define amdgpu_kernel void @v_extract_fabs_fold_v2f16(ptr addrspace(1) %in) #0 {
; CI-LABEL: v_extract_fabs_fold_v2f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; CI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_add_i32_e32 v0, vcc, s0, v0
; CI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT: flat_load_dword v0, v[0:1]
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: v_lshrrev_b32_e32 v1, 16, v0
; CI-NEXT: v_cvt_f32_f16_e64 v0, |v0|
; CI-NEXT: v_cvt_f32_f16_e64 v1, |v1|
; CI-NEXT: v_mul_f32_e32 v0, 4.0, v0
; CI-NEXT: v_add_f32_e32 v1, 2.0, v1
; CI-NEXT: v_cvt_f16_f32_e32 v0, v0
; CI-NEXT: v_cvt_f16_f32_e32 v1, v1
; CI-NEXT: flat_store_short v[0:1], v0
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: flat_store_short v[0:1], v1
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: s_endpgm
;
; VI-LABEL: v_extract_fabs_fold_v2f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; VI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: v_add_u32_e32 v0, vcc, s0, v0
; VI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT: flat_load_dword v0, v[0:1]
; VI-NEXT: v_mov_b32_e32 v1, 0x4000
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: v_mul_f16_e64 v2, |v0|, 4.0
; VI-NEXT: v_add_f16_sdwa v0, |v0|, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; VI-NEXT: flat_store_short v[0:1], v2
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: flat_store_short v[0:1], v0
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: v_extract_fabs_fold_v2f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: v_mov_b32_e32 v1, 0x4000
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: global_load_dword v0, v0, s[0:1]
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_mul_f16_e64 v2, |v0|, 4.0
; GFX9-NEXT: v_add_f16_sdwa v0, |v0|, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; GFX9-NEXT: global_store_short v[0:1], v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: global_store_short v[0:1], v0, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.in = getelementptr inbounds <2 x half>, ptr addrspace(1) %in, i32 %tid
%val = load <2 x half>, ptr addrspace(1) %gep.in
%fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %val)
%elt0 = extractelement <2 x half> %fabs, i32 0
%elt1 = extractelement <2 x half> %fabs, i32 1
%fmul0 = fmul half %elt0, 4.0
%fadd1 = fadd half %elt1, 2.0
store volatile half %fmul0, ptr addrspace(1) undef
store volatile half %fadd1, ptr addrspace(1) undef
ret void
}
define amdgpu_kernel void @v_extract_fabs_no_fold_v2f16(ptr addrspace(1) %in) #0 {
; CI-LABEL: v_extract_fabs_no_fold_v2f16:
; CI: ; %bb.0:
; CI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; CI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT: s_waitcnt lgkmcnt(0)
; CI-NEXT: v_mov_b32_e32 v1, s1
; CI-NEXT: v_add_i32_e32 v0, vcc, s0, v0
; CI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT: flat_load_dword v0, v[0:1]
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: v_bfe_u32 v1, v0, 16, 15
; CI-NEXT: v_and_b32_e32 v0, 0x7fff, v0
; CI-NEXT: flat_store_short v[0:1], v0
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: flat_store_short v[0:1], v1
; CI-NEXT: s_waitcnt vmcnt(0)
; CI-NEXT: s_endpgm
;
; VI-LABEL: v_extract_fabs_no_fold_v2f16:
; VI: ; %bb.0:
; VI-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; VI-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; VI-NEXT: s_waitcnt lgkmcnt(0)
; VI-NEXT: v_mov_b32_e32 v1, s1
; VI-NEXT: v_add_u32_e32 v0, vcc, s0, v0
; VI-NEXT: v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT: flat_load_dword v0, v[0:1]
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: v_and_b32_e32 v1, 0x7fff7fff, v0
; VI-NEXT: v_bfe_u32 v0, v0, 16, 15
; VI-NEXT: flat_store_short v[0:1], v1
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: flat_store_short v[0:1], v0
; VI-NEXT: s_waitcnt vmcnt(0)
; VI-NEXT: s_endpgm
;
; GFX9-LABEL: v_extract_fabs_no_fold_v2f16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x0
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: global_load_dword v0, v0, s[0:1]
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_and_b32_e32 v0, 0x7fff7fff, v0
; GFX9-NEXT: global_store_short v[0:1], v0, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: global_store_short_d16_hi v[0:1], v0, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.in = getelementptr inbounds <2 x half>, ptr addrspace(1) %in, i32 %tid
%val = load <2 x half>, ptr addrspace(1) %gep.in
%fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %val)
%elt0 = extractelement <2 x half> %fabs, i32 0
%elt1 = extractelement <2 x half> %fabs, i32 1
store volatile half %elt0, ptr addrspace(1) undef
store volatile half %elt1, ptr addrspace(1) undef
ret void
}
declare half @llvm.fabs.f16(half) #1
declare <2 x half> @llvm.fabs.v2f16(<2 x half>) #1
declare <4 x half> @llvm.fabs.v4f16(<4 x half>) #1
declare i32 @llvm.amdgcn.workitem.id.x() #1
attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; GCN: {{.*}}
; GFX89: {{.*}}
Reference in New Issue View Git Blame Copy Permalink