mirror of
https://github.com/RPCSX/llvm.git
synced 2024-12-25 13:35:39 +00:00
7eba65d30c
For some reason there are both of these available, except for scalar 64-bit compares which only has u64. I'm not sure why there are both (I'm guessing it's for the one bit inputs we don't use), but for consistency always using the unsigned one. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@282832 91177308-0d34-0410-b5e6-96231b3b80d8
140 lines
4.9 KiB
LLVM
140 lines
4.9 KiB
LLVM
; RUN: llc -march=amdgcn -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
|
|
; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
|
|
; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=R600 -check-prefix=FUNC %s
|
|
|
|
; FUNC-LABEL: {{^}}s_uint_to_fp_i32_to_f32:
|
|
; SI: v_cvt_f32_u32_e32
|
|
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW]}}, KC0[2].Z
|
|
define void @s_uint_to_fp_i32_to_f32(float addrspace(1)* %out, i32 %in) #0 {
|
|
%result = uitofp i32 %in to float
|
|
store float %result, float addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}v_uint_to_fp_i32_to_f32:
|
|
; SI: v_cvt_f32_u32_e32 {{v[0-9]+}}, {{v[0-9]+$}}
|
|
|
|
; R600: INT_TO_FLT
|
|
define void @v_uint_to_fp_i32_to_f32(float addrspace(1)* %out, i32 addrspace(1)* %in) #0 {
|
|
%tid = call i32 @llvm.r600.read.tidig.x()
|
|
%in.gep = getelementptr i32, i32 addrspace(1)* %in, i32 %tid
|
|
%out.gep = getelementptr float, float addrspace(1)* %out, i32 %tid
|
|
%val = load i32, i32 addrspace(1)* %in.gep
|
|
%result = uitofp i32 %val to float
|
|
store float %result, float addrspace(1)* %out.gep
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}s_uint_to_fp_v2i32_to_v2f32:
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: v_cvt_f32_u32_e32
|
|
|
|
; R600-DAG: UINT_TO_FLT * T{{[0-9]+\.[XYZW]}}, KC0[2].W
|
|
; R600-DAG: UINT_TO_FLT * T{{[0-9]+\.[XYZW]}}, KC0[3].X
|
|
define void @s_uint_to_fp_v2i32_to_v2f32(<2 x float> addrspace(1)* %out, <2 x i32> %in) #0 {
|
|
%result = uitofp <2 x i32> %in to <2 x float>
|
|
store <2 x float> %result, <2 x float> addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}s_uint_to_fp_v4i32_to_v4f32:
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: s_endpgm
|
|
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
define void @s_uint_to_fp_v4i32_to_v4f32(<4 x float> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) #0 {
|
|
%value = load <4 x i32>, <4 x i32> addrspace(1) * %in
|
|
%result = uitofp <4 x i32> %value to <4 x float>
|
|
store <4 x float> %result, <4 x float> addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}v_uint_to_fp_v4i32:
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: v_cvt_f32_u32_e32
|
|
; SI: v_cvt_f32_u32_e32
|
|
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; R600: UINT_TO_FLT * T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
define void @v_uint_to_fp_v4i32(<4 x float> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) #0 {
|
|
%tid = call i32 @llvm.r600.read.tidig.x()
|
|
%in.gep = getelementptr <4 x i32>, <4 x i32> addrspace(1)* %in, i32 %tid
|
|
%out.gep = getelementptr <4 x float>, <4 x float> addrspace(1)* %out, i32 %tid
|
|
%value = load <4 x i32>, <4 x i32> addrspace(1)* %in.gep
|
|
%result = uitofp <4 x i32> %value to <4 x float>
|
|
store <4 x float> %result, <4 x float> addrspace(1)* %out.gep
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}s_uint_to_fp_i1_to_f32:
|
|
; SI: v_cmp_eq_u32_e64 [[CMP:s\[[0-9]+:[0-9]\]]],
|
|
; SI-NEXT: v_cndmask_b32_e64 [[RESULT:v[0-9]+]], 0, 1.0, [[CMP]]
|
|
; SI: buffer_store_dword [[RESULT]],
|
|
; SI: s_endpgm
|
|
define void @s_uint_to_fp_i1_to_f32(float addrspace(1)* %out, i32 %in) #0 {
|
|
%cmp = icmp eq i32 %in, 0
|
|
%fp = uitofp i1 %cmp to float
|
|
store float %fp, float addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}s_uint_to_fp_i1_to_f32_load:
|
|
; SI: v_cndmask_b32_e64 [[RESULT:v[0-9]+]], 0, 1.0
|
|
; SI: buffer_store_dword [[RESULT]],
|
|
; SI: s_endpgm
|
|
define void @s_uint_to_fp_i1_to_f32_load(float addrspace(1)* %out, i1 %in) #0 {
|
|
%fp = uitofp i1 %in to float
|
|
store float %fp, float addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}v_uint_to_fp_i1_f32_load:
|
|
; SI: {{buffer|flat}}_load_ubyte
|
|
; SI: v_and_b32_e32 {{v[0-9]+}}, 1, {{v[0-9]+}}
|
|
; SI: v_cmp_eq_u32
|
|
; SI: v_cndmask_b32_e64 [[RESULT:v[0-9]+]], 0, 1.0
|
|
; SI: {{buffer|flat}}_store_dword {{.*}}[[RESULT]]
|
|
; SI: s_endpgm
|
|
define void @v_uint_to_fp_i1_f32_load(float addrspace(1)* %out, i1 addrspace(1)* %in) #0 {
|
|
%tid = call i32 @llvm.r600.read.tidig.x()
|
|
%in.gep = getelementptr i1, i1 addrspace(1)* %in, i32 %tid
|
|
%out.gep = getelementptr float, float addrspace(1)* %out, i32 %tid
|
|
%val = load i1, i1 addrspace(1)* %in.gep
|
|
%fp = uitofp i1 %val to float
|
|
store float %fp, float addrspace(1)* %out.gep
|
|
ret void
|
|
}
|
|
|
|
; FIXME: Repeated here to test r600
|
|
; FUNC-LABEL: {{^}}s_uint_to_fp_i64_to_f32:
|
|
; R600: FFBH_UINT
|
|
; R600: FFBH_UINT
|
|
; R600: CNDE_INT
|
|
; R600: CNDE_INT
|
|
|
|
; R600-DAG: SETGT_UINT
|
|
; R600-DAG: SETGT_UINT
|
|
; R600-DAG: SETE_INT
|
|
|
|
define void @s_uint_to_fp_i64_to_f32(float addrspace(1)* %out, i64 %in) #0 {
|
|
entry:
|
|
%cvt = uitofp i64 %in to float
|
|
store float %cvt, float addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
declare i32 @llvm.r600.read.tidig.x() #1
|
|
|
|
attributes #0 = { nounwind }
|
|
attributes #1 = { nounwind readnone }
|