1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 03:02:36 +01:00
llvm-mirror/test/CodeGen/AMDGPU/add.v2i16.ll
Matt Arsenault ee27f88e14 AMDGPU: Make v2i16/v2f16 legal on VI
This usually results in better code. Fixes using
inline asm with short2, and also fixes having a different
ABI for function parameters between VI and gfx9.

Partially cleans up the mess used for lowering of the d16
operations. Making v4f16 legal will help clean this up more,
but this requires additional work.

llvm-svn: 332953
2018-05-22 06:32:10 +00:00

287 lines
13 KiB
LLVM

; RUN: llc -march=amdgcn -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GFX9,GCN %s
; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,VI %s
; FIXME: Need to handle non-uniform case for function below (load without gep).
; GCN-LABEL: {{^}}v_test_add_v2i16:
; GFX9: v_pk_add_u16 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}
; FIXME: or should be unnecessary
; VI: v_add_u16_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}
; VI: v_add_u16_sdwa v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; VI: v_or_b32
define amdgpu_kernel void @v_test_add_v2i16(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0, <2 x i16> addrspace(1)* %in1) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%gep.in1 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in1, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%b = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in1
%add = add <2 x i16> %a, %b
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; GCN-LABEL: {{^}}s_test_add_v2i16:
; GFX9: s_load_dword [[VAL0:s[0-9]+]]
; GFX9: s_load_dword [[VAL1:s[0-9]+]]
; GFX9: v_mov_b32_e32 [[VVAL1:v[0-9]+]]
; GFX9: v_pk_add_u16 v{{[0-9]+}}, [[VAL0]], [[VVAL1]]
; VI: s_add_i32
; VI: s_add_i32
define amdgpu_kernel void @s_test_add_v2i16(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(4)* %in0, <2 x i16> addrspace(4)* %in1) #1 {
%a = load <2 x i16>, <2 x i16> addrspace(4)* %in0
%b = load <2 x i16>, <2 x i16> addrspace(4)* %in1
%add = add <2 x i16> %a, %b
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; GCN-LABEL: {{^}}s_test_add_self_v2i16:
; GFX9: s_load_dword [[VAL:s[0-9]+]]
; GFX9: v_pk_add_u16 v{{[0-9]+}}, [[VAL]], [[VAL]]
; VI: s_add_i32
; VI: s_add_i32
define amdgpu_kernel void @s_test_add_self_v2i16(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(4)* %in0) #1 {
%a = load <2 x i16>, <2 x i16> addrspace(4)* %in0
%add = add <2 x i16> %a, %a
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; FIXME: VI should not scalarize arg access.
; GCN-LABEL: {{^}}s_test_add_v2i16_kernarg:
; GFX9: v_pk_add_u16 v{{[0-9]+}}, s{{[0-9]+}}, v{{[0-9]+}}
; VI: s_add_i32
; VI: s_add_i32
; VI: s_lshl_b32 s{{[0-9]+}}, s{{[0-9]+}}, 16
; VI: s_and_b32
; VI: s_or_b32
define amdgpu_kernel void @s_test_add_v2i16_kernarg(<2 x i16> addrspace(1)* %out, <2 x i16> %a, <2 x i16> %b) #1 {
%add = add <2 x i16> %a, %b
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; FIXME: Eliminate or with sdwa
; GCN-LABEL: {{^}}v_test_add_v2i16_constant:
; GFX9: s_mov_b32 [[CONST:s[0-9]+]], 0x1c8007b{{$}}
; GFX9: v_pk_add_u16 v{{[0-9]+}}, v{{[0-9]+}}, [[CONST]]
; VI-DAG: v_mov_b32_e32 v[[SCONST:[0-9]+]], 0x1c8
; VI-DAG: v_add_u16_e32 v{{[0-9]+}}, 0x7b, v{{[0-9]+}}
; VI-DAG: v_add_u16_sdwa v{{[0-9]+}}, v{{[0-9]+}}, v[[SCONST]] dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; VI: v_or_b32_e32
define amdgpu_kernel void @v_test_add_v2i16_constant(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%add = add <2 x i16> %a, <i16 123, i16 456>
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; FIXME: Need to handle non-uniform case for function below (load without gep).
; GCN-LABEL: {{^}}v_test_add_v2i16_neg_constant:
; GFX9: s_mov_b32 [[CONST:s[0-9]+]], 0xfc21fcb3{{$}}
; GFX9: v_pk_add_u16 v{{[0-9]+}}, v{{[0-9]+}}, [[CONST]]
; VI-DAG: v_add_u16_e32 v{{[0-9]+}}, 0xfffffcb3, v{{[0-9]+}}
; VI-DAG: v_mov_b32_e32 v[[SCONST:[0-9]+]], 0xfffffc21
; VI-DAG: v_add_u16_sdwa v{{[0-9]+}}, v{{[0-9]+}}, v[[SCONST]] dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
define amdgpu_kernel void @v_test_add_v2i16_neg_constant(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%add = add <2 x i16> %a, <i16 -845, i16 -991>
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; GCN-LABEL: {{^}}v_test_add_v2i16_inline_neg1:
; GFX9: v_pk_add_u16 v{{[0-9]+}}, v{{[0-9]+}}, -1 op_sel_hi:[1,0]{{$}}
; VI: v_mov_b32_e32 v[[SCONST:[0-9]+]], -1
; VI: flat_load_dword [[LOAD:v[0-9]+]]
; VI-DAG: v_add_u16_sdwa v{{[0-9]+}}, [[LOAD]], v[[SCONST]] dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; VI-DAG: v_add_u16_e32 v{{[0-9]+}}, -1, [[LOAD]]
; VI: v_or_b32_e32
define amdgpu_kernel void @v_test_add_v2i16_inline_neg1(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%add = add <2 x i16> %a, <i16 -1, i16 -1>
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; GCN-LABEL: {{^}}v_test_add_v2i16_inline_lo_zero_hi:
; GFX9: v_pk_add_u16 v{{[0-9]+}}, v{{[0-9]+}}, 32{{$}}
; VI: flat_load_dword
; VI-NOT: v_add_u16
; VI: v_and_b32_e32 v{{[0-9]+}}, 0xffff0000,
; VI: v_add_u16_e32 v{{[0-9]+}}, 32, v{{[0-9]+}}
; VI-NOT: v_add_u16
; VI: v_or_b32_e32
define amdgpu_kernel void @v_test_add_v2i16_inline_lo_zero_hi(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%add = add <2 x i16> %a, <i16 32, i16 0>
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; The high element gives fp
; GCN-LABEL: {{^}}v_test_add_v2i16_inline_fp_split:
; GFX9: s_mov_b32 [[K:s[0-9]+]], 1.0
; GFX9: v_pk_add_u16 v{{[0-9]+}}, v{{[0-9]+}}, [[K]]{{$}}
; VI-NOT: v_add_u16
; VI: v_mov_b32_e32 v[[K:[0-9]+]], 0x3f80
; VI: v_add_u16_sdwa v{{[0-9]+}}, v{{[0-9]+}}, v[[K]] dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; VI-NOT: v_add_u16
; VI: v_or_b32_sdwa v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_0 src1_sel:DWORD
define amdgpu_kernel void @v_test_add_v2i16_inline_fp_split(<2 x i16> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%add = add <2 x i16> %a, <i16 0, i16 16256>
store <2 x i16> %add, <2 x i16> addrspace(1)* %out
ret void
}
; FIXME: Need to handle non-uniform case for function below (load without gep).
; GCN-LABEL: {{^}}v_test_add_v2i16_zext_to_v2i32:
; GFX9: global_load_dword [[A:v[0-9]+]]
; GFX9: global_load_dword [[B:v[0-9]+]]
; GFX9: v_pk_add_u16 [[ADD:v[0-9]+]], [[A]], [[B]]
; GFX9-DAG: v_and_b32_e32 v[[ELT0:[0-9]+]], 0xffff, [[ADD]]
; GFX9-DAG: v_lshrrev_b32_e32 v[[ELT1:[0-9]+]], 16, [[ADD]]
; GFX9: buffer_store_dwordx2 v{{\[}}[[ELT0]]:[[ELT1]]{{\]}}
; VI: flat_load_dword v[[A:[0-9]+]]
; VI: flat_load_dword v[[B:[0-9]+]]
; VI-NOT: and
; VI-NOT: shl
; VI: v_add_u16_e32 v[[ADD_LO:[0-9]+]], v[[A]], v[[B]]
; VI: v_add_u16_sdwa v[[ADD_HI:[0-9]+]], v[[A]], v[[B]] dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; VI-NOT: and
; VI-NOT: shl
; VI: buffer_store_dwordx2 v{{\[}}[[ADD_LO]]:[[ADD_HI]]{{\]}}
define amdgpu_kernel void @v_test_add_v2i16_zext_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0, <2 x i16> addrspace(1)* %in1) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i32>, <2 x i32> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%gep.in1 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in1, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%b = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in1
%add = add <2 x i16> %a, %b
%ext = zext <2 x i16> %add to <2 x i32>
store <2 x i32> %ext, <2 x i32> addrspace(1)* %out
ret void
}
; FIXME: Need to handle non-uniform case for function below (load without gep).
; GCN-LABEL: {{^}}v_test_add_v2i16_zext_to_v2i64:
; GFX9: global_load_dword [[A:v[0-9]+]]
; GFX9: global_load_dword [[B:v[0-9]+]]
; GFX9: v_pk_add_u16 [[ADD:v[0-9]+]], [[A]], [[B]]
; GFX9-DAG: v_and_b32_e32 v[[ELT0:[0-9]+]], 0xffff, [[ADD]]
; GFX9-DAG: v_lshrrev_b32_e32 v[[ELT1:[0-9]+]], 16, [[ADD]]
; GFX9: buffer_store_dwordx4
; VI-DAG: v_mov_b32_e32 v{{[0-9]+}}, 0{{$}}
; VI: flat_load_dword v[[A:[0-9]+]]
; VI: flat_load_dword v[[B:[0-9]+]]
; VI-DAG: v_add_u16_e32
; VI: v_add_u16_sdwa v[[ADD_HI:[0-9]+]], v[[A]], v[[B]] dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; VI: buffer_store_dwordx4
define amdgpu_kernel void @v_test_add_v2i16_zext_to_v2i64(<2 x i64> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0, <2 x i16> addrspace(1)* %in1) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i64>, <2 x i64> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%gep.in1 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in1, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%b = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in1
%add = add <2 x i16> %a, %b
%ext = zext <2 x i16> %add to <2 x i64>
store <2 x i64> %ext, <2 x i64> addrspace(1)* %out
ret void
}
; FIXME: Need to handle non-uniform case for function below (load without gep).
; GCN-LABEL: {{^}}v_test_add_v2i16_sext_to_v2i32:
; GFX9: global_load_dword [[A:v[0-9]+]]
; GFX9: global_load_dword [[B:v[0-9]+]]
; GFX9: v_pk_add_u16 [[ADD:v[0-9]+]], [[A]], [[B]]
; GFX9-DAG: v_bfe_i32 v[[ELT0:[0-9]+]], [[ADD]], 0, 16
; GFX9-DAG: v_ashrrev_i32_e32 v[[ELT1:[0-9]+]], 16, [[ADD]]
; GFX9: buffer_store_dwordx2 v{{\[}}[[ELT0]]:[[ELT1]]{{\]}}
; VI: v_add_u16_sdwa v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}} dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; VI: v_add_u16_e32
; VI: v_bfe_i32 v{{[0-9]+}}, v{{[0-9]+}}, 0, 16
; VI: v_bfe_i32 v{{[0-9]+}}, v{{[0-9]+}}, 0, 16
; VI: buffer_store_dwordx2
define amdgpu_kernel void @v_test_add_v2i16_sext_to_v2i32(<2 x i32> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0, <2 x i16> addrspace(1)* %in1) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i32>, <2 x i32> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%gep.in1 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in1, i32 %tid
%a = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%b = load volatile <2 x i16>, <2 x i16> addrspace(1)* %gep.in1
%add = add <2 x i16> %a, %b
%ext = sext <2 x i16> %add to <2 x i32>
store <2 x i32> %ext, <2 x i32> addrspace(1)* %out
ret void
}
; FIXME: Need to handle non-uniform case for function below (load without gep).
; GCN-LABEL: {{^}}v_test_add_v2i16_sext_to_v2i64:
; GCN: {{flat|global}}_load_dword
; GCN: {{flat|global}}_load_dword
; GFX9: v_pk_add_u16
; GFX9: v_lshrrev_b32_e32 v{{[0-9]+}}, 16, v{{[0-9]+}}
; VI: v_add_u16_sdwa
; VI: v_add_u16_e32
; GCN: v_bfe_i32 v{{[0-9]+}}, v{{[0-9]+}}, 0, 16
; GCN: v_bfe_i32 v{{[0-9]+}}, v{{[0-9]+}}, 0, 16
; GCN: v_ashrrev_i32_e32 v{{[0-9]+}}, 31, v{{[0-9]+}}
; GCN: v_ashrrev_i32_e32 v{{[0-9]+}}, 31, v{{[0-9]+}}
define amdgpu_kernel void @v_test_add_v2i16_sext_to_v2i64(<2 x i64> addrspace(1)* %out, <2 x i16> addrspace(1)* %in0, <2 x i16> addrspace(1)* %in1) #1 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%gep.out = getelementptr inbounds <2 x i64>, <2 x i64> addrspace(1)* %out, i32 %tid
%gep.in0 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in0, i32 %tid
%gep.in1 = getelementptr inbounds <2 x i16>, <2 x i16> addrspace(1)* %in1, i32 %tid
%a = load <2 x i16>, <2 x i16> addrspace(1)* %gep.in0
%b = load <2 x i16>, <2 x i16> addrspace(1)* %gep.in1
%add = add <2 x i16> %a, %b
%ext = sext <2 x i16> %add to <2 x i64>
store <2 x i64> %ext, <2 x i64> addrspace(1)* %out
ret void
}
declare i32 @llvm.amdgcn.workitem.id.x() #0
attributes #0 = { nounwind readnone }
attributes #1 = { nounwind }