dbf278b984
Form the MAI spec: It’s ok that Src_C and vDst are the exact same VGPRs or Src_C and vDst are completely separated. The case that Src_C and vDst are overlapping should be avoid as new value could be written to accumulator input before it gets read. Note that this inevitably increases register pressure to the point where some programs will become uncompilable. This patch separates MAC and FMA versions of MFMA instructions using either tied dst and src2 or earlyclobber dst. Fixes: SWDEV-318900 Differential Revision: https://reviews.llvm.org/D117844
128 lines
6.5 KiB
LLVM
128 lines
6.5 KiB
LLVM
; RUN: llc -march=amdgcn -mcpu=gfx908 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX908 %s
|
|
; RUN: llc -march=amdgcn -mcpu=gfx90a -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX90A %s
|
|
|
|
; GCN-LABEL: {{^}}max_12regs_13a_used:
|
|
; GCN-NOT: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD0
|
|
; GCN-NOT: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD1
|
|
; GCN: v_accvgpr_read_b32 v[[VSPILL:[0-9]+]], a{{[0-9]+}}
|
|
; GCN-NOT: buffer_store_dword
|
|
; GCN-NOT: buffer_load_dword
|
|
; GCN: v_accvgpr_write_b32 a{{[0-9]+}}, v[[VSPILL]]
|
|
; GCN: ScratchSize: 0
|
|
define amdgpu_kernel void @max_12regs_13a_used(i32 %cond, <4 x float> addrspace(1)* %arg, <4 x float> addrspace(1)* %out) #2 {
|
|
bb:
|
|
%in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
|
|
%mai.1 = tail call <4 x float> @llvm.amdgcn.mfma.f32.4x4x1f32(float 1.0, float 1.0, <4 x float> %in.1, i32 0, i32 0, i32 0)
|
|
%mai.2 = tail call <4 x float> @llvm.amdgcn.mfma.f32.4x4x1f32(float 1.0, float 1.0, <4 x float> %mai.1, i32 0, i32 0, i32 0)
|
|
%cmp = icmp eq i32 %cond, 0
|
|
br i1 %cmp, label %use, label %st
|
|
|
|
use:
|
|
call void asm sideeffect "", "a,a,a,a,a"(i32 1, i32 2, i32 3, i32 4, i32 5)
|
|
store volatile <4 x float> <float 1.0, float 1.0, float 1.0, float 1.0>, <4 x float> addrspace(1)* %out
|
|
br label %st
|
|
|
|
st:
|
|
%gep1 = getelementptr <4 x float>, <4 x float> addrspace(1)* %out, i64 16
|
|
%gep2 = getelementptr <4 x float>, <4 x float> addrspace(1)* %out, i64 32
|
|
call void asm sideeffect "", "a,a"(<4 x float> %mai.1, <4 x float> %mai.2)
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}max_10_vgprs_used_9a:
|
|
; GCN-NOT: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD0
|
|
; GCN-NOT: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD1
|
|
; GCN: v_accvgpr_read_b32 v[[VSPILL:[0-9]+]], a{{[0-9]+}}
|
|
; GCN-NOT: buffer_store_dword
|
|
; GCN-NOT: buffer_load_dword
|
|
; GCN: v_accvgpr_write_b32 a{{[0-9]+}}, v[[VSPILL]]
|
|
; GCN: ScratchSize: 0
|
|
define amdgpu_kernel void @max_10_vgprs_used_9a() #1 {
|
|
%a1 = call <4 x i32> asm sideeffect "", "=a"()
|
|
%a2 = call <4 x i32> asm sideeffect "", "=a"()
|
|
%a3 = call i32 asm sideeffect "", "=a"()
|
|
%a4 = call <2 x i32> asm sideeffect "", "=a"()
|
|
call void asm sideeffect "", "a,a,a"(<4 x i32> %a1, <4 x i32> %a2, i32 %a3)
|
|
call void asm sideeffect "", "a"(<2 x i32> %a4)
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}max_32regs_mfma32:
|
|
; GCN-NOT: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD0
|
|
; GCN-NOT: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD1
|
|
; GCN-NOT: buffer_store_dword
|
|
; GCN: v_accvgpr_read_b32
|
|
; GCN: v_mfma_f32_32x32x1f32
|
|
; GCN-NOT: buffer_load_dword
|
|
; GCN: v_accvgpr_write_b32
|
|
; GCN: ScratchSize: 0
|
|
define amdgpu_kernel void @max_32regs_mfma32(float addrspace(1)* %arg) #3 {
|
|
bb:
|
|
%v = call i32 asm sideeffect "", "=a"()
|
|
br label %use
|
|
|
|
use:
|
|
%mai.1 = tail call <32 x float> @llvm.amdgcn.mfma.f32.32x32x1f32(float 1.0, float 1.0, <32 x float> <float 1.0, float 2.0, float 3.0, float 4.0, float 5.0, float 6.0, float 7.0, float 8.0, float 9.0, float 10.0, float 11.0, float 12.0, float 13.0, float 14.0, float 15.0, float 16.0, float 17.0, float 18.0, float 19.0, float 20.0, float 21.0, float 22.0, float 23.0, float 24.0, float 25.0, float 26.0, float 27.0, float 28.0, float 29.0, float 30.0, float 31.0, float 2.0>, i32 0, i32 0, i32 0)
|
|
call void asm sideeffect "", "a"(i32 %v)
|
|
%elt1 = extractelement <32 x float> %mai.1, i32 0
|
|
store float %elt1, float addrspace(1)* %arg
|
|
ret void
|
|
}
|
|
|
|
; Should spill agprs to memory for both gfx908 and gfx90a.
|
|
; GCN-LABEL: {{^}}max_5regs_used_8a:
|
|
; GCN: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD0
|
|
; GCN: s_mov_b32 s{{[0-9]+}}, SCRATCH_RSRC_DWORD1
|
|
|
|
; GFX908-DAG: v_accvgpr_read_b32 v1, a0 ; Reload Reuse
|
|
; GFX908-DAG: buffer_store_dword v1, off, s[{{[0-9:]+}}], 0 offset:4 ; 4-byte Folded Spill
|
|
; GFX908-DAG: v_accvgpr_read_b32 v1, a1 ; Reload Reuse
|
|
; GFX908-DAG: buffer_store_dword v1, off, s[{{[0-9:]+}}], 0 offset:8 ; 4-byte Folded Spill
|
|
; GFX908-DAG: v_accvgpr_read_b32 v1, a2 ; Reload Reuse
|
|
; GFX908-DAG: buffer_store_dword v1, off, s[{{[0-9:]+}}], 0 offset:12 ; 4-byte Folded Spill
|
|
; GFX908-DAG: v_accvgpr_read_b32 v1, a3 ; Reload Reuse
|
|
; GFX908-DAG: buffer_store_dword v1, off, s[{{[0-9:]+}}], 0 offset:16 ; 4-byte Folded Spill
|
|
|
|
; GFX90A-DAG: buffer_store_dword a0, off, s[{{[0-9:]+}}], 0 offset:4 ; 4-byte Folded Spill
|
|
; GFX90A-DAG: buffer_store_dword a1, off, s[{{[0-9:]+}}], 0 offset:8 ; 4-byte Folded Spill
|
|
; GFX90A-DAG: v_accvgpr_read_b32 v4, a2 ; Reload Reuse
|
|
; GFX90A-DAG: v_accvgpr_read_b32 v3, a3 ; Reload Reuse
|
|
|
|
; GCN: v_mfma_f32_4x4x1f32 a[0:3], v{{[0-9]+}}, v{{[0-9]+}}, a[0:3]
|
|
|
|
; GFX908-DAG: buffer_load_dword v0, off, s[{{[0-9:]+}}], 0 offset:4 ; 4-byte Folded Reload
|
|
; GFX908-DAG: buffer_load_dword v1, off, s[{{[0-9:]+}}], 0 offset:8 ; 4-byte Folded Reload
|
|
; GFX908-DAG: buffer_load_dword v2, off, s[{{[0-9:]+}}], 0 offset:12 ; 4-byte Folded Reload
|
|
; GFX908-DAG: buffer_load_dword v3, off, s[{{[0-9:]+}}], 0 offset:16 ; 4-byte Folded Reload
|
|
; GFX908: global_store_dwordx4 v[{{[0-9:]+}}], v[0:3], off
|
|
|
|
; GFX90A-DAG: buffer_load_dword a0, off, s[4:7], 0 offset:4 ; 4-byte Folded Reload
|
|
; GFX90A-DAG: buffer_load_dword a1, off, s[4:7], 0 offset:8 ; 4-byte Folded Reload
|
|
; GFX90A-DAG: v_accvgpr_write_b32 a2, v4 ; Reload Reuse
|
|
; GFX90A-DAG: v_accvgpr_write_b32 a3, v3 ; Reload Reuse
|
|
; GFX90A: global_store_dwordx4 v[0:1], a[0:3], off
|
|
|
|
; GCN: ScratchSize: 20
|
|
define amdgpu_kernel void @max_5regs_used_8a(<4 x float> addrspace(1)* %arg) #4 {
|
|
%tid = call i32 @llvm.amdgcn.workitem.id.x()
|
|
%v0 = call float asm sideeffect "; def $0", "=v"()
|
|
%a4 = call <4 x float> asm sideeffect "; def $0", "=a"()
|
|
%gep = getelementptr inbounds <4 x float>, <4 x float> addrspace(1)* %arg, i32 %tid
|
|
%mai.in = load <4 x float>, <4 x float> addrspace(1)* %gep
|
|
%mai.out = tail call <4 x float> @llvm.amdgcn.mfma.f32.4x4x1f32(float 1.0, float 1.0, <4 x float> %mai.in, i32 0, i32 0, i32 0)
|
|
store <4 x float> %mai.out, <4 x float> addrspace(1)* %gep
|
|
store volatile <4 x float> %a4, <4 x float> addrspace(1)* undef
|
|
call void asm sideeffect "; use $0", "v"(float %v0);
|
|
ret void
|
|
}
|
|
|
|
declare i32 @llvm.amdgcn.workitem.id.x()
|
|
declare <16 x float> @llvm.amdgcn.mfma.f32.16x16x1f32(float, float, <16 x float>, i32, i32, i32)
|
|
declare <4 x float> @llvm.amdgcn.mfma.f32.4x4x1f32(float, float, <4 x float>, i32, i32, i32)
|
|
declare <32 x float> @llvm.amdgcn.mfma.f32.32x32x1f32(float, float, <32 x float>, i32, i32, i32)
|
|
|
|
attributes #1 = { nounwind "amdgpu-num-vgpr"="10" }
|
|
attributes #2 = { nounwind "amdgpu-num-vgpr"="12" }
|
|
attributes #3 = { nounwind "amdgpu-num-vgpr"="32" }
|
|
attributes #4 = { nounwind "amdgpu-num-vgpr"="5" }
|