1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 19:52:54 +01:00
llvm-mirror/test/CodeGen/AMDGPU/llvm.amdgcn.fmul.legacy.ll
Stanislav Mekhanoshin e23fa40f7f [AMDGPU] Allow SDWA in instructions with immediates and SGPRs
An encoding does not allow to use SDWA in an instruction with
scalar operands, either literals or SGPRs. That is however possible
to copy these operands into a VGPR first.

Several copies of the value are produced if multiple SDWA conversions
were done. To cleanup MachineLICM (to hoist copies out of loops),
MachineCSE (to remove duplicate copies) and SIFoldOperands (to replace
SGPR to VGPR copy with immediate copy right to the VGPR) runs are added
after the SDWA pass.

Differential Revision: https://reviews.llvm.org/D33583

llvm-svn: 304219
2017-05-30 16:49:24 +00:00

55 lines
2.2 KiB
LLVM

; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
; GCN-LABEL: {{^}}test_mul_legacy_f32:
; GCN: v_mul_legacy_f32_e32 v{{[0-9]+}}, s{{[0-9]+}}, v{{[0-9]+}}
define amdgpu_kernel void @test_mul_legacy_f32(float addrspace(1)* %out, float %a, float %b) #0 {
%result = call float @llvm.amdgcn.fmul.legacy(float %a, float %b)
store float %result, float addrspace(1)* %out, align 4
ret void
}
; GCN-LABEL: {{^}}test_mul_legacy_undef0_f32:
; GCN: v_mul_legacy_f32_e32
define amdgpu_kernel void @test_mul_legacy_undef0_f32(float addrspace(1)* %out, float %a) #0 {
%result = call float @llvm.amdgcn.fmul.legacy(float undef, float %a)
store float %result, float addrspace(1)* %out, align 4
ret void
}
; GCN-LABEL: {{^}}test_mul_legacy_undef1_f32:
; GCN: v_mul_legacy_f32_e32
define amdgpu_kernel void @test_mul_legacy_undef1_f32(float addrspace(1)* %out, float %a) #0 {
%result = call float @llvm.amdgcn.fmul.legacy(float %a, float undef)
store float %result, float addrspace(1)* %out, align 4
ret void
}
; GCN-LABEL: {{^}}test_mul_legacy_fabs_f32:
; GCN: v_mul_legacy_f32_e64 v{{[0-9]+}}, |s{{[0-9]+}}|, |v{{[0-9]+}}|
define amdgpu_kernel void @test_mul_legacy_fabs_f32(float addrspace(1)* %out, float %a, float %b) #0 {
%a.fabs = call float @llvm.fabs.f32(float %a)
%b.fabs = call float @llvm.fabs.f32(float %b)
%result = call float @llvm.amdgcn.fmul.legacy(float %a.fabs, float %b.fabs)
store float %result, float addrspace(1)* %out, align 4
ret void
}
; TODO: Should match mac_legacy/mad_legacy
; GCN-LABEL: {{^}}test_mad_legacy_f32:
; GCN: v_mul_legacy_f32_e32 v{{[0-9]+}}, s{{[0-9]+}}, v{{[0-9]+}}
; GCN: v_add_f32_e32
define amdgpu_kernel void @test_mad_legacy_f32(float addrspace(1)* %out, float %a, float %b, float %c) #0 {
%mul = call float @llvm.amdgcn.fmul.legacy(float %a, float %b)
%add = fadd float %mul, %c
store float %add, float addrspace(1)* %out, align 4
ret void
}
declare float @llvm.fabs.f32(float) #1
declare float @llvm.amdgcn.fmul.legacy(float, float) #1
attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }