; RUN: opt -S -mtriple=amdgcn- -passes=sroa %s -o %t.sroa.ll
; RUN: llc -mtriple=amdgcn-- -mcpu=tonga -mattr=-promote-alloca -verify-machineinstrs < %t.sroa.ll | FileCheck -enable-var-scope --check-prefixes=GCN,GCN-ALLOCA %s
; RUN: llc -mtriple=amdgcn-- -mcpu=tonga -mattr=+promote-alloca -verify-machineinstrs < %t.sroa.ll | FileCheck -enable-var-scope --check-prefixes=GCN,GCN-PROMOTE %s
; RUN: opt -S -mtriple=amdgcn-- -passes='sroa,amdgpu-promote-alloca,instcombine' < %s | FileCheck -check-prefix=OPT %s
target datalayout = "A5"
; OPT-LABEL: @vector_read_alloca_bitcast(
; OPT-NOT: alloca
; OPT: %0 = extractelement <4 x i32> <i32 0, i32 1, i32 2, i32 3>, i32 %index
; OPT-NEXT: store i32 %0, ptr addrspace(1) %out, align 4
; GCN-LABEL: {{^}}vector_read_alloca_bitcast:
; GCN-ALLOCA-COUNT-4: buffer_store_dword
; GCN-ALLOCA: buffer_load_dword
; GCN-PROMOTE: s_cmp_eq_u32 s{{[0-9]+}}, 1
; GCN-PROMOTE: s_cselect_b64 [[CC1:[^,]+]], -1, 0
; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 2
; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND1:v[0-9]+]], 0, 1, [[CC1]]
; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0
; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 3
; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND2:v[0-9]+]], 2, [[IND1]], vcc
; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0
; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND3:v[0-9]+]], 3, [[IND2]], vcc
; GCN-PROMOTE: ScratchSize: 0
define amdgpu_kernel void @vector_read_alloca_bitcast(ptr addrspace(1) %out, i32 %index) {
entry:
%tmp = alloca [4 x i32], addrspace(5)
%y = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 1
%z = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 2
%w = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 3
store i32 0, ptr addrspace(5) %tmp
store i32 1, ptr addrspace(5) %y
store i32 2, ptr addrspace(5) %z
store i32 3, ptr addrspace(5) %w
%tmp1 = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 %index
%tmp2 = load i32, ptr addrspace(5) %tmp1
store i32 %tmp2, ptr addrspace(1) %out
ret void
}
; OPT-LABEL: @vector_write_alloca_bitcast(
; OPT-NOT: alloca
; OPT: %0 = insertelement <4 x i32> zeroinitializer, i32 1, i32 %w_index
; OPT-NEXT: %1 = extractelement <4 x i32> %0, i32 %r_index
; OPT-NEXT: store i32 %1, ptr addrspace(1) %out, align
; GCN-LABEL: {{^}}vector_write_alloca_bitcast:
; GCN-ALLOCA-COUNT-5: buffer_store_dword
; GCN-ALLOCA: buffer_load_dword
; GCN-PROMOTE-COUNT-7: v_cndmask
; GCN-PROMOTE: ScratchSize: 0
define amdgpu_kernel void @vector_write_alloca_bitcast(ptr addrspace(1) %out, i32 %w_index, i32 %r_index) {
entry:
%tmp = alloca [4 x i32], addrspace(5)
%y = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 1
%z = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 2
%w = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 3
store i32 0, ptr addrspace(5) %tmp
store i32 0, ptr addrspace(5) %y
store i32 0, ptr addrspace(5) %z
store i32 0, ptr addrspace(5) %w
%tmp1 = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 %w_index
store i32 1, ptr addrspace(5) %tmp1
%tmp2 = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 %r_index
%tmp3 = load i32, ptr addrspace(5) %tmp2
store i32 %tmp3, ptr addrspace(1) %out
ret void
}
; OPT-LABEL: @vector_write_read_bitcast_to_float(
; OPT-NOT: alloca
; OPT: bb2:
; OPT: %promotealloca = phi <6 x float> [ undef, %bb ], [ %0, %bb2 ]
; OPT: %0 = insertelement <6 x float> %promotealloca, float %tmp71, i32 %tmp10
; OPT: .preheader:
; OPT: %bc = bitcast <6 x float> %0 to <6 x i32>
; OPT: %1 = extractelement <6 x i32> %bc, i32 %tmp20
; GCN-LABEL: {{^}}vector_write_read_bitcast_to_float:
; GCN-ALLOCA: buffer_store_dword
; GCN-PROMOTE: v_cmp_eq_u16
; GCN-PROMOTE: v_cndmask
; GCN: s_cbranch
; GCN-ALLOCA: buffer_load_dword
; GCN-PROMOTE: ScratchSize: 0
define amdgpu_kernel void @vector_write_read_bitcast_to_float(ptr addrspace(1) %arg) {
bb:
%tmp = alloca [6 x float], align 4, addrspace(5)
call void @llvm.lifetime.start.p5(i64 24, ptr addrspace(5) %tmp) #2
br label %bb2
bb2: ; preds = %bb2, %bb
%tmp3 = phi i32 [ 0, %bb ], [ %tmp13, %bb2 ]
%tmp4 = zext i32 %tmp3 to i64
%tmp5 = getelementptr inbounds float, ptr addrspace(1) %arg, i64 %tmp4
%tmp7 = load i32, ptr addrspace(1) %tmp5, align 4
%tmp8 = trunc i32 %tmp3 to i16
%tmp9 = urem i16 %tmp8, 6
%tmp10 = zext i16 %tmp9 to i32
%tmp11 = getelementptr inbounds [6 x float], ptr addrspace(5) %tmp, i32 0, i32 %tmp10
store i32 %tmp7, ptr addrspace(5) %tmp11, align 4
%tmp13 = add nuw nsw i32 %tmp3, 1
%tmp14 = icmp eq i32 %tmp13, 1000
br i1 %tmp14, label %.preheader, label %bb2
bb15: ; preds = %.preheader
call void @llvm.lifetime.end.p5(i64 24, ptr addrspace(5) %tmp) #2
ret void
.preheader: ; preds = %.preheader, %bb2
%tmp16 = phi i32 [ %tmp27, %.preheader ], [ 0, %bb2 ]
%tmp17 = trunc i32 %tmp16 to i16
%tmp18 = urem i16 %tmp17, 6
%tmp19 = sub nuw nsw i16 5, %tmp18
%tmp20 = zext i16 %tmp19 to i32
%tmp21 = getelementptr inbounds [6 x float], ptr addrspace(5) %tmp, i32 0, i32 %tmp20
%tmp23 = load i32, ptr addrspace(5) %tmp21, align 4
%tmp24 = zext i32 %tmp16 to i64
%tmp25 = getelementptr inbounds float, ptr addrspace(1) %arg, i64 %tmp24
store i32 %tmp23, ptr addrspace(1) %tmp25, align 4
%tmp27 = add nuw nsw i32 %tmp16, 1
%tmp28 = icmp eq i32 %tmp27, 1000
br i1 %tmp28, label %bb15, label %.preheader
}
; OPT-LABEL: @vector_write_read_bitcast_to_double(
; OPT-NOT: alloca
; OPT: bb2:
; OPT: %promotealloca = phi <6 x double> [ undef, %bb ], [ %0, %bb2 ]
; OPT: %0 = insertelement <6 x double> %promotealloca, double %tmp71, i32 %tmp10
; OPT: .preheader:
; OPT: %bc = bitcast <6 x double> %0 to <6 x i64>
; OPT: %1 = extractelement <6 x i64> %bc, i32 %tmp20
; GCN-LABEL: {{^}}vector_write_read_bitcast_to_double:
; GCN-ALLOCA-COUNT-2: buffer_store_dword
; GCN-PROMOTE-COUNT-2: v_movreld_b32_e32
; GCN: s_cbranch
; GCN-ALLOCA-COUNT-2: buffer_load_dword
; GCN-PROMOTE-COUNT-2: v_movrels_b32_e32
; GCN-PROMOTE: ScratchSize: 0
define amdgpu_kernel void @vector_write_read_bitcast_to_double(ptr addrspace(1) %arg) {
bb:
%tmp = alloca [6 x double], align 8, addrspace(5)
call void @llvm.lifetime.start.p5(i64 48, ptr addrspace(5) %tmp) #2
br label %bb2
bb2: ; preds = %bb2, %bb
%tmp3 = phi i32 [ 0, %bb ], [ %tmp13, %bb2 ]
%tmp4 = zext i32 %tmp3 to i64
%tmp5 = getelementptr inbounds double, ptr addrspace(1) %arg, i64 %tmp4
%tmp7 = load i64, ptr addrspace(1) %tmp5, align 8
%tmp8 = trunc i32 %tmp3 to i16
%tmp9 = urem i16 %tmp8, 6
%tmp10 = zext i16 %tmp9 to i32
%tmp11 = getelementptr inbounds [6 x double], ptr addrspace(5) %tmp, i32 0, i32 %tmp10
store i64 %tmp7, ptr addrspace(5) %tmp11, align 8
%tmp13 = add nuw nsw i32 %tmp3, 1
%tmp14 = icmp eq i32 %tmp13, 1000
br i1 %tmp14, label %.preheader, label %bb2
bb15: ; preds = %.preheader
call void @llvm.lifetime.end.p5(i64 48, ptr addrspace(5) %tmp) #2
ret void
.preheader: ; preds = %.preheader, %bb2
%tmp16 = phi i32 [ %tmp27, %.preheader ], [ 0, %bb2 ]
%tmp17 = trunc i32 %tmp16 to i16
%tmp18 = urem i16 %tmp17, 6
%tmp19 = sub nuw nsw i16 5, %tmp18
%tmp20 = zext i16 %tmp19 to i32
%tmp21 = getelementptr inbounds [6 x double], ptr addrspace(5) %tmp, i32 0, i32 %tmp20
%tmp23 = load i64, ptr addrspace(5) %tmp21, align 8
%tmp24 = zext i32 %tmp16 to i64
%tmp25 = getelementptr inbounds double, ptr addrspace(1) %arg, i64 %tmp24
store i64 %tmp23, ptr addrspace(1) %tmp25, align 8
%tmp27 = add nuw nsw i32 %tmp16, 1
%tmp28 = icmp eq i32 %tmp27, 1000
br i1 %tmp28, label %bb15, label %.preheader
}
; OPT-LABEL: @vector_write_read_bitcast_to_i64(
; OPT-NOT: alloca
; OPT: bb2:
; OPT: %promotealloca = phi <6 x i64> [ undef, %bb ], [ %0, %bb2 ]
; OPT: %0 = insertelement <6 x i64> %promotealloca, i64 %tmp6, i32 %tmp9
; OPT: .preheader:
; OPT: %1 = extractelement <6 x i64> %0, i32 %tmp18
; GCN-LABEL: {{^}}vector_write_read_bitcast_to_i64:
; GCN-ALLOCA-COUNT-2: buffer_store_dword
; GCN-PROMOTE-COUNT-2: v_movreld_b32_e32
; GCN: s_cbranch
; GCN-ALLOCA-COUNT-2: buffer_load_dword
; GCN-PROMOTE-COUNT-2: v_movrels_b32_e32
; GCN-PROMOTE: ScratchSize: 0
define amdgpu_kernel void @vector_write_read_bitcast_to_i64(ptr addrspace(1) %arg) {
bb:
%tmp = alloca [6 x i64], align 8, addrspace(5)
call void @llvm.lifetime.start.p5(i64 48, ptr addrspace(5) %tmp) #2
br label %bb2
bb2: ; preds = %bb2, %bb
%tmp3 = phi i32 [ 0, %bb ], [ %tmp11, %bb2 ]
%tmp4 = zext i32 %tmp3 to i64
%tmp5 = getelementptr inbounds i64, ptr addrspace(1) %arg, i64 %tmp4
%tmp6 = load i64, ptr addrspace(1) %tmp5, align 8
%tmp7 = trunc i32 %tmp3 to i16
%tmp8 = urem i16 %tmp7, 6
%tmp9 = zext i16 %tmp8 to i32
%tmp10 = getelementptr inbounds [6 x i64], ptr addrspace(5) %tmp, i32 0, i32 %tmp9
store i64 %tmp6, ptr addrspace(5) %tmp10, align 8
%tmp11 = add nuw nsw i32 %tmp3, 1
%tmp12 = icmp eq i32 %tmp11, 1000
br i1 %tmp12, label %.preheader, label %bb2
bb13: ; preds = %.preheader
call void @llvm.lifetime.end.p5(i64 48, ptr addrspace(5) %tmp) #2
ret void
.preheader: ; preds = %.preheader, %bb2
%tmp14 = phi i32 [ %tmp23, %.preheader ], [ 0, %bb2 ]
%tmp15 = trunc i32 %tmp14 to i16
%tmp16 = urem i16 %tmp15, 6
%tmp17 = sub nuw nsw i16 5, %tmp16
%tmp18 = zext i16 %tmp17 to i32
%tmp19 = getelementptr inbounds [6 x i64], ptr addrspace(5) %tmp, i32 0, i32 %tmp18
%tmp20 = load i64, ptr addrspace(5) %tmp19, align 8
%tmp21 = zext i32 %tmp14 to i64
%tmp22 = getelementptr inbounds i64, ptr addrspace(1) %arg, i64 %tmp21
store i64 %tmp20, ptr addrspace(1) %tmp22, align 8
%tmp23 = add nuw nsw i32 %tmp14, 1
%tmp24 = icmp eq i32 %tmp23, 1000
br i1 %tmp24, label %bb13, label %.preheader
}
; TODO: llvm.assume can be ingored
; OPT-LABEL: @vector_read_alloca_bitcast_assume(
; OPT: %0 = extractelement <4 x i32> <i32 0, i32 1, i32 2, i32 3>, i32 %index
; OPT: store i32 %0, ptr addrspace(1) %out, align 4
; GCN-LABEL: {{^}}vector_read_alloca_bitcast_assume:
; GCN-COUNT: buffer_store_dword
define amdgpu_kernel void @vector_read_alloca_bitcast_assume(ptr addrspace(1) %out, i32 %index) {
entry:
%tmp = alloca [4 x i32], addrspace(5)
%cmp = icmp ne ptr addrspace(5) %tmp, null
call void @llvm.assume(i1 %cmp)
%y = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 1
%z = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 2
%w = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 3
store i32 0, ptr addrspace(5) %tmp
store i32 1, ptr addrspace(5) %y
store i32 2, ptr addrspace(5) %z
store i32 3, ptr addrspace(5) %w
%tmp1 = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 %index
%tmp2 = load i32, ptr addrspace(5) %tmp1
store i32 %tmp2, ptr addrspace(1) %out
ret void
}
; OPT-LABEL: @vector_read_alloca_multiuse(
; OPT-NOT: alloca
; OPT: %0 = extractelement <4 x i32> <i32 0, i32 1, i32 2, i32 3>, i32 %index
; OPT-NEXT: %add2 = add nuw nsw i32 %0, 1
; OPT-NEXT: store i32 %add2, ptr addrspace(1) %out, align 4
; GCN-LABEL: {{^}}vector_read_alloca_multiuse:
; GCN-ALLOCA-COUNT-4: buffer_store_dword
; GCN-ALLOCA: buffer_load_dword
; GCN-PROMOTE: s_cmp_eq_u32 s{{[0-9]+}}, 1
; GCN-PROMOTE: s_cselect_b64 [[CC1:[^,]+]], -1, 0
; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 2
; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND1:v[0-9]+]], 0, 1, [[CC1]]
; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0
; GCN-PROMOTE: s_cmp_lg_u32 s{{[0-9]+}}, 3
; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND2:v[0-9]+]], 2, [[IND1]], vcc
; GCN-PROMOTE: s_cselect_b64 vcc, -1, 0
; GCN-PROMOTE: v_cndmask_b32_e{{32|64}} [[IND3:v[0-9]+]], 3, [[IND2]], vcc
; GCN-PROMOTE: ScratchSize: 0
define amdgpu_kernel void @vector_read_alloca_multiuse(ptr addrspace(1) %out, i32 %index) {
entry:
%tmp = alloca [4 x i32], addrspace(5)
%y = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 1
%z = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 2
%w = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 3
store i32 0, ptr addrspace(5) %tmp
store i32 1, ptr addrspace(5) %y
store i32 2, ptr addrspace(5) %z
store i32 3, ptr addrspace(5) %w
%tmp1 = getelementptr [4 x i32], ptr addrspace(5) %tmp, i32 0, i32 %index
%tmp2 = load i32, ptr addrspace(5) %tmp1
%tmp3 = load i32, ptr addrspace(5) %tmp
%tmp4 = load i32, ptr addrspace(5) %y
%add1 = add i32 %tmp2, %tmp3
%add2 = add i32 %add1, %tmp4
store i32 %add2, ptr addrspace(1) %out
ret void
}
; OPT-LABEL: @bitcast_vector_to_vector(
; OPT-NOT: alloca
; OPT: store <4 x i32> <i32 1, i32 2, i32 3, i32 4>, ptr addrspace(1) %out, align 16
; GCN-LABEL: {{^}}bitcast_vector_to_vector:
; GCN: v_mov_b32_e32 v0, 1
; GCN: v_mov_b32_e32 v1, 2
; GCN: v_mov_b32_e32 v2, 3
; GCN: v_mov_b32_e32 v3, 4
; GCN: ScratchSize: 0
define amdgpu_kernel void @bitcast_vector_to_vector(ptr addrspace(1) %out) {
.entry:
%alloca = alloca <4 x float>, align 16, addrspace(5)
store <4 x i32> <i32 1, i32 2, i32 3, i32 4>, ptr addrspace(5) %alloca
%load = load <4 x i32>, ptr addrspace(5) %alloca, align 16
store <4 x i32> %load, ptr addrspace(1) %out
ret void
}
; OPT-LABEL: @vector_bitcast_from_alloca_array(
; OPT-NOT: alloca
; OPT: store <4 x i32> <i32 1, i32 2, i32 3, i32 4>, ptr addrspace(1) %out, align 16
; GCN-LABEL: {{^}}vector_bitcast_from_alloca_array:
; GCN: v_mov_b32_e32 v0, 1
; GCN: v_mov_b32_e32 v1, 2
; GCN: v_mov_b32_e32 v2, 3
; GCN: v_mov_b32_e32 v3, 4
; GCN: ScratchSize: 0
define amdgpu_kernel void @vector_bitcast_from_alloca_array(ptr addrspace(1) %out) {
.entry:
%alloca = alloca [4 x float], align 16, addrspace(5)
store <4 x i32> <i32 1, i32 2, i32 3, i32 4>, ptr addrspace(5) %alloca
%load = load <4 x i32>, ptr addrspace(5) %alloca, align 16
store <4 x i32> %load, ptr addrspace(1) %out
ret void
}
; OPT-LABEL: @vector_bitcast_to_array_from_alloca_array(
; OPT-NOT: alloca
; OPT-NEXT: store i32 1, ptr addrspace(1) %out, align 4
; OPT-NEXT: %out.repack1 = getelementptr inbounds i8, ptr addrspace(1) %out, i64 4
; OPT-NEXT: store i32 2, ptr addrspace(1) %out.repack1, align 4
; OPT-NEXT: %out.repack2 = getelementptr inbounds i8, ptr addrspace(1) %out, i64 8
; OPT-NEXT: store i32 3, ptr addrspace(1) %out.repack2, align 4
; OPT-NEXT: %out.repack3 = getelementptr inbounds i8, ptr addrspace(1) %out, i64 12
; OPT-NEXT: store i32 4, ptr addrspace(1) %out.repack3, align 4
; GCN-LABEL: {{^}}vector_bitcast_to_array_from_alloca_array:
; GCN: v_mov_b32_e32 v0, 1
; GCN: v_mov_b32_e32 v1, 2
; GCN: v_mov_b32_e32 v2, 3
; GCN: v_mov_b32_e32 v3, 4
; GCN: ScratchSize: 0
define amdgpu_kernel void @vector_bitcast_to_array_from_alloca_array(ptr addrspace(1) %out) {
%alloca = alloca [4 x float], align 16, addrspace(5)
store [4 x i32] [i32 1, i32 2, i32 3, i32 4], ptr addrspace(5) %alloca
%load = load [4 x i32], ptr addrspace(5) %alloca, align 16
store [4 x i32] %load, ptr addrspace(1) %out
ret void
}
; OPT-LABEL: @vector_bitcast_to_struct_from_alloca_array(
; OPT-NOT: alloca
; OPT-NEXT: store i32 1, ptr addrspace(1) %out, align 4
; OPT-NEXT: %out.repack1 = getelementptr inbounds i8, ptr addrspace(1) %out, i64 4
; OPT-NEXT: store i32 2, ptr addrspace(1) %out.repack1, align 4
; OPT-NEXT: %out.repack2 = getelementptr inbounds i8, ptr addrspace(1) %out, i64 8
; OPT-NEXT: store i32 3, ptr addrspace(1) %out.repack2, align 4
; OPT-NEXT: %out.repack3 = getelementptr inbounds i8, ptr addrspace(1) %out, i64 12
; OPT-NEXT: store i32 4, ptr addrspace(1) %out.repack3, align 4
; GCN-LABEL: {{^}}vector_bitcast_to_struct_from_alloca_array:
; GCN: v_mov_b32_e32 v0, 1
; GCN: v_mov_b32_e32 v1, 2
; GCN: v_mov_b32_e32 v2, 3
; GCN: v_mov_b32_e32 v3, 4
; GCN: ScratchSize: 0
%struct.v4 = type { i32, i32, i32, i32 }
define amdgpu_kernel void @vector_bitcast_to_struct_from_alloca_array(ptr addrspace(1) %out) {
%alloca = alloca [4 x float], align 16, addrspace(5)
store %struct.v4 { i32 1, i32 2, i32 3, i32 4 }, ptr addrspace(5) %alloca
%load = load %struct.v4, ptr addrspace(5) %alloca, align 16
store %struct.v4 %load, ptr addrspace(1) %out
ret void
}
declare void @llvm.lifetime.start.p5(i64 immarg, ptr addrspace(5) nocapture)
declare void @llvm.lifetime.end.p5(i64 immarg, ptr addrspace(5) nocapture)
declare void @llvm.assume(i1)
Codebase Browser
llvm