1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
|
;=========================== begin_copyright_notice ============================
;
; Copyright (C) 2021 Intel Corporation
;
; SPDX-License-Identifier: MIT
;
;============================ end_copyright_notice =============================
; RUN: %opt %use_old_pass_manager% -GenXPrologEpilogInsertion -dce -mattr=+ocl_runtime -march=genx64 \
; RUN: -mtriple=spir64-unknown-unknown -mcpu=Gen9 -S -vc-arg-reg-size=32 \
; RUN: -vc-ret-reg-size=12 < %s | FileCheck %s
target datalayout = "e-p:64:64-i64:64-n8:16:32:64"
target triple = "genx64-unknown-unknown"
declare <8 x float> @llvm.genx.oword.ld.unaligned.v8f32(i32, i32, i32)
declare void @llvm.genx.oword.st.v8f32(i32, i32, <8 x float>)
define dllexport void @test(i32 %arg, i32 %x0, i32 %x1, i32 %x2, i32 %x3, i32 %x4, i32 %x5, i32 %x6, i64 %x7, i64 %x8) local_unnamed_addr #0 {
entry:
%a0 = tail call <8 x float> @llvm.genx.oword.ld.unaligned.v8f32(i32 0, i32 1, i32 %arg)
%a1 = tail call <8 x float> @llvm.genx.oword.ld.unaligned.v8f32(i32 0, i32 2, i32 %arg)
%a2 = tail call <8 x float> @llvm.genx.oword.ld.unaligned.v8f32(i32 0, i32 3, i32 %arg)
%a3 = tail call <8 x float> @llvm.genx.oword.ld.unaligned.v8f32(i32 0, i32 4, i32 %arg)
; call is below: foo(undef, a0, a1, a2, a3, a2)
; from first %a2 (correct part)
; %14 = call <8 x float> @llvm.genx.write.predef.reg.v8f32.v256f32(i32 8, <256 x float> %13)
; %15 = call <256 x float> @llvm.genx.read.predef.reg.v256f32.v256f32(i32 8, <256 x float> undef)
; %16 = call <256 x float> @llvm.genx.wrregionf.v256f32.v8f32.i16.i1(<256 x float> %15, <8 x float> %a2, i32 0, i32 8, i32 1, i16 96, i32 undef, i1 true)
; %17 = call <8 x float> @llvm.genx.write.predef.reg.v8f32.v256f32(i32 8, <256 x float> %16)
; CHECK: %[[wrreg:[^ ]+]] = call <256 x float> @llvm.genx.wrregionf.v256f32.v8f32.i16.i1(<256 x float> %{{[^ ]+}}, <8 x float> %a2
; CHECK: %[[wrongsubst:[^ ]+]] = call <8 x float> @llvm.genx.write.predef.reg.v8f32.v256f32(i32 8, <256 x float> %[[wrreg]])
; and then reusing this %17 as an argument
; from repeated %a2 (wrong part)
; %21 = call <256 x float> @llvm.genx.read.predef.reg.v256f32.v256f32(i32 8, <256 x float> undef)
; %22 = call <256 x float> @llvm.genx.wrregionf.v256f32.v8f32.i16.i1(<256 x float> %21, <8 x float> %17, i32 0, i32 8, i32 1, i16 160, i32 undef, i1 true)
; %23 = call <8 x float> @llvm.genx.write.predef.reg.v8f32.v256f32(i32 8, <256 x float> %22)
; CHECK-NOT: call <256 x float> @llvm.genx.wrregionf.v256f32.v8f32.i16.i1(<256 x float> %{{[^ ]+}}, <8 x float> %[[wrongsubst]]
; really we need one more slot formed from %a2 (correct thing)
; CHECK: call <256 x float> @llvm.genx.wrregionf.v256f32.v8f32.i16.i1(<256 x float> %{{[^ ]+}}, <8 x float> %a2
; this is incorrect because in <8 x float> %17 we are loosing information, we need <8 x float> %a2 instead
%ret = tail call spir_func <8 x float> @foo(<8 x float> undef, <8 x float> %a0, <8 x float> %a1, <8 x float> %a2, <8 x float> %a3, <8 x float> %a2)
tail call void @llvm.genx.oword.st.v8f32(i32 5, i32 %arg, <8 x float> %ret)
ret void
}
define internal spir_func <8 x float> @foo(<8 x float> %0, <8 x float> %1, <8 x float> %2, <8 x float> %3, <8 x float> %4, <8 x float> %5) #1 {
entry:
%add.i = fadd <8 x float> %1, %2
%add.i7 = fadd <8 x float> %add.i, %3
%add.i13 = fadd <8 x float> %add.i7, %4
%add.i19 = fadd <8 x float> %add.i13, %5
ret <8 x float> %add.i19
}
attributes #0 = { nounwind "CMGenxMain" "oclrt"="1" }
attributes #1 = { noinline norecurse nounwind readnone "CMStackCall" }
;!spirv.Source = !{!0}
;!opencl.spir.version = !{!1}
;!opencl.ocl.version = !{!2}
;!opencl.used.extensions = !{!3}
;!opencl.used.optional.core.features = !{!3}
;!spirv.Generator = !{!4}
!genx.kernels = !{!5}
!genx.kernel.internal = !{!15}
!0 = !{i32 0, i32 100000}
!1 = !{i32 1, i32 2}
!2 = !{i32 1, i32 0}
!3 = !{}
!4 = !{i16 6, i16 14}
!5 = !{void (i32, i32, i32, i32, i32, i32, i32, i32, i64, i64)* @test, !"test", !6, i32 0, !7, !8, !9, i32 0}
!6 = !{i32 112, i32 2, i32 2, i32 2, i32 2, i32 2, i32 24, i32 8, i32 96, i32 104}
!7 = !{i32 -1, i32 160, i32 168, i32 176, i32 184, i32 192, i32 64, i32 128, i32 144, i32 152}
!8 = !{i32 0, i32 0, i32 0, i32 0, i32 0, i32 0}
!9 = !{!"svmptr_t", !"buffer_t", !"buffer_t", !"buffer_t", !"buffer_t", !"buffer_t"}
!15 = !{void (i32, i32, i32, i32, i32, i32, i32, i32, i64, i64)* @test, !16, !17, !18, !22}
!16 = !{i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0}
!17 = !{i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 0}
!18 = !{!19}
!19 = !{i32 0, !20}
!20 = !{!21}
!21 = !{i32 9, i32 0}
!22 = !{i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9}
|
