RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-22 20:43:44 +02:00
Code Issues Projects Releases Wiki Activity
805c8ee060
llvm-mirror/test/CodeGen/PowerPC/ppc64-P9-vabsd.ll
Hiroshi Inoue 0a8b8fc8aa [PowerPC] fix incorrect vectorization of abs() on POWER9 
Vectorized loops with abs() returns incorrect results on POWER9. This patch fixes it.
For example the following code returns negative result if input values are negative though it sums up the absolute value of the inputs.

int vpx_satd_c(const int16_t *coeff, int length) {
  int satd = 0;
  for (int i = 0; i < length; ++i) satd += abs(coeff[i]);
  return satd;
}

This problem causes test failures for libvpx.
For vector absolute and vector absolute difference on POWER9, LLVM generates VABSDUW (Vector Absolute Difference Unsigned Word) instruction or variants.
Since these instructions are for unsigned integers, we need adjustment for signed integers.
For abs(sub(a, b)), we generate VABSDUW(a+0x80000000, b+0x80000000). Otherwise, abs(sub(-1, 0)) returns 0xFFFFFFFF(=-1) instead of 1. For abs(a), we generate VABSDUW(a+0x80000000, 0x80000000).

Differential Revision: https://reviews.llvm.org/D45522

llvm-svn: 330497
2018-04-21 09:32:17 +00:00

367 lines
13 KiB
LLVM
Raw Blame History

; RUN: llc < %s -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr9 -verify-machineinstrs | FileCheck %s
; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr9 -verify-machineinstrs | FileCheck %s
; RUN: llc < %s -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr8 -verify-machineinstrs | FileCheck %s -check-prefix=CHECK-PWR8 -implicit-check-not vabsdu
; Function Attrs: nounwind readnone
define <4 x i32> @simple_absv_32(<4 x i32> %a) local_unnamed_addr {
entry:
%sub.i = sub <4 x i32> zeroinitializer, %a
%0 = tail call <4 x i32> @llvm.ppc.altivec.vmaxsw(<4 x i32> %a, <4 x i32> %sub.i)
ret <4 x i32> %0
; CHECK-LABEL: simple_absv_32
; CHECK-DAG: vxor {{[0-9]+}}, [[REG:[0-9]+]], [[REG]]
; CHECK-DAG: xvnegsp 34, 34
; CHECK-DAG: xvnegsp 35, {{[0-9]+}}
; CHECK-NEXT: vabsduw 2, 2, {{[0-9]+}}
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: simple_absv_32
; CHECK-PWR8: xxlxor
; CHECK-PWR8: vsubuwm
; CHECK-PWR8: vmaxsw
; CHECK-PWR8: blr
}
; Function Attrs: nounwind readnone
define <4 x i32> @simple_absv_32_swap(<4 x i32> %a) local_unnamed_addr {
entry:
%sub.i = sub <4 x i32> zeroinitializer, %a
%0 = tail call <4 x i32> @llvm.ppc.altivec.vmaxsw(<4 x i32> %sub.i, <4 x i32> %a)
ret <4 x i32> %0
; CHECK-LABEL: simple_absv_32_swap
; CHECK-DAG: vxor {{[0-9]+}}, [[REG:[0-9]+]], [[REG]]
; CHECK-DAG: xvnegsp 34, 34
; CHECK-DAG: xvnegsp 35, {{[0-9]+}}
; CHECK-NEXT: vabsduw 2, 2, {{[0-9]+}}
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: simple_absv_32_swap
; CHECK-PWR8: xxlxor
; CHECK-PWR8: vsubuwm
; CHECK-PWR8: vmaxsw
; CHECK-PWR8: blr
}
define <8 x i16> @simple_absv_16(<8 x i16> %a) local_unnamed_addr {
entry:
%sub.i = sub <8 x i16> zeroinitializer, %a
%0 = tail call <8 x i16> @llvm.ppc.altivec.vmaxsh(<8 x i16> %a, <8 x i16> %sub.i)
ret <8 x i16> %0
; CHECK-LABEL: simple_absv_16
; CHECK: mtvsrws {{[0-9]+}}, {{[0-9]+}}
; CHECK-NEXT: vadduhm 2, 2, [[IMM:[0-9]+]]
; CHECK-NEXT: vabsduh 2, 2, [[IMM]]
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: simple_absv_16
; CHECK-PWR8: xxlxor
; CHECK-PWR8: vsubuhm
; CHECK-PWR8: vmaxsh
; CHECK-PWR8: blr
}
; Function Attrs: nounwind readnone
define <16 x i8> @simple_absv_8(<16 x i8> %a) local_unnamed_addr {
entry:
%sub.i = sub <16 x i8> zeroinitializer, %a
%0 = tail call <16 x i8> @llvm.ppc.altivec.vmaxsb(<16 x i8> %a, <16 x i8> %sub.i)
ret <16 x i8> %0
; CHECK-LABEL: simple_absv_8
; CHECK: xxspltib {{[0-9]+}}, 128
; CHECK-NEXT: vaddubm 2, 2, [[IMM:[0-9]+]]
; CHECK-NEXT: vabsdub 2, 2, [[IMM]]
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: simple_absv_8
; CHECK-PWR8: xxlxor
; CHECK-PWR8: vsububm
; CHECK-PWR8: vmaxsb
; CHECK-PWR8: blr
}
; The select pattern can only be detected for v4i32.
; Function Attrs: norecurse nounwind readnone
define <4 x i32> @sub_absv_32(<4 x i32> %a, <4 x i32> %b) local_unnamed_addr {
entry:
%0 = sub nsw <4 x i32> %a, %b
%1 = icmp sgt <4 x i32> %0, <i32 -1, i32 -1, i32 -1, i32 -1>
%2 = sub <4 x i32> zeroinitializer, %0
%3 = select <4 x i1> %1, <4 x i32> %0, <4 x i32> %2
ret <4 x i32> %3
; CHECK-LABEL: sub_absv_32
; CHECK-DAG: xvnegsp 34, 34
; CHECK-DAG: xvnegsp 35, 35
; CHECK-NEXT: vabsduw 2, 2, 3
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: sub_absv_32
; CHECK-PWR8: vsubuwm
; CHECK-PWR8: xxlxor
; CHECK-PWR8: blr
}
; FIXME: This does not produce the ISD::ABS that we are looking for.
; We should fix the missing canonicalization.
; We do manage to find the word version of ABS but not the halfword.
; Threfore, we end up doing more work than is required with a pair of abs for word
; instead of just one for the halfword.
; Function Attrs: norecurse nounwind readnone
define <8 x i16> @sub_absv_16(<8 x i16> %a, <8 x i16> %b) local_unnamed_addr {
entry:
%0 = sext <8 x i16> %a to <8 x i32>
%1 = sext <8 x i16> %b to <8 x i32>
%2 = sub nsw <8 x i32> %0, %1
%3 = icmp sgt <8 x i32> %2, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
%4 = sub nsw <8 x i32> zeroinitializer, %2
%5 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> %4
%6 = trunc <8 x i32> %5 to <8 x i16>
ret <8 x i16> %6
; CHECK-LABEL: sub_absv_16
; CHECK-NOT: vabsduh
; CHECK: vabsduw
; CHECK-NOT: vabsduh
; CHECK: vabsduw
; CHECK-NOT: vabsduh
; CHECK: blr
; CHECK-PWR8-LABEL: sub_absv_16
; CHECK-PWR8: vsubuwm
; CHECK-PWR8: xxlxor
; CHECK-PWR8: blr
}
; FIXME: This does not produce ISD::ABS. This does not even vectorize correctly!
; This function should look like sub_absv_32 and sub_absv_16 except that the type is v16i8.
; Function Attrs: norecurse nounwind readnone
define <16 x i8> @sub_absv_8(<16 x i8> %a, <16 x i8> %b) local_unnamed_addr {
entry:
%vecext = extractelement <16 x i8> %a, i32 0
%conv = zext i8 %vecext to i32
%vecext1 = extractelement <16 x i8> %b, i32 0
%conv2 = zext i8 %vecext1 to i32
%sub = sub nsw i32 %conv, %conv2
%ispos = icmp sgt i32 %sub, -1
%neg = sub nsw i32 0, %sub
%0 = select i1 %ispos, i32 %sub, i32 %neg
%conv3 = trunc i32 %0 to i8
%vecins = insertelement <16 x i8> <i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>, i8 %conv3, i32 0
%vecext4 = extractelement <16 x i8> %a, i32 1
%conv5 = zext i8 %vecext4 to i32
%vecext6 = extractelement <16 x i8> %b, i32 1
%conv7 = zext i8 %vecext6 to i32
%sub8 = sub nsw i32 %conv5, %conv7
%ispos171 = icmp sgt i32 %sub8, -1
%neg172 = sub nsw i32 0, %sub8
%1 = select i1 %ispos171, i32 %sub8, i32 %neg172
%conv10 = trunc i32 %1 to i8
%vecins11 = insertelement <16 x i8> %vecins, i8 %conv10, i32 1
%vecext12 = extractelement <16 x i8> %a, i32 2
%conv13 = zext i8 %vecext12 to i32
%vecext14 = extractelement <16 x i8> %b, i32 2
%conv15 = zext i8 %vecext14 to i32
%sub16 = sub nsw i32 %conv13, %conv15
%ispos173 = icmp sgt i32 %sub16, -1
%neg174 = sub nsw i32 0, %sub16
%2 = select i1 %ispos173, i32 %sub16, i32 %neg174
%conv18 = trunc i32 %2 to i8
%vecins19 = insertelement <16 x i8> %vecins11, i8 %conv18, i32 2
%vecext20 = extractelement <16 x i8> %a, i32 3
%conv21 = zext i8 %vecext20 to i32
%vecext22 = extractelement <16 x i8> %b, i32 3
%conv23 = zext i8 %vecext22 to i32
%sub24 = sub nsw i32 %conv21, %conv23
%ispos175 = icmp sgt i32 %sub24, -1
%neg176 = sub nsw i32 0, %sub24
%3 = select i1 %ispos175, i32 %sub24, i32 %neg176
%conv26 = trunc i32 %3 to i8
%vecins27 = insertelement <16 x i8> %vecins19, i8 %conv26, i32 3
%vecext28 = extractelement <16 x i8> %a, i32 4
%conv29 = zext i8 %vecext28 to i32
%vecext30 = extractelement <16 x i8> %b, i32 4
%conv31 = zext i8 %vecext30 to i32
%sub32 = sub nsw i32 %conv29, %conv31
%ispos177 = icmp sgt i32 %sub32, -1
%neg178 = sub nsw i32 0, %sub32
%4 = select i1 %ispos177, i32 %sub32, i32 %neg178
%conv34 = trunc i32 %4 to i8
%vecins35 = insertelement <16 x i8> %vecins27, i8 %conv34, i32 4
%vecext36 = extractelement <16 x i8> %a, i32 5
%conv37 = zext i8 %vecext36 to i32
%vecext38 = extractelement <16 x i8> %b, i32 5
%conv39 = zext i8 %vecext38 to i32
%sub40 = sub nsw i32 %conv37, %conv39
%ispos179 = icmp sgt i32 %sub40, -1
%neg180 = sub nsw i32 0, %sub40
%5 = select i1 %ispos179, i32 %sub40, i32 %neg180
%conv42 = trunc i32 %5 to i8
%vecins43 = insertelement <16 x i8> %vecins35, i8 %conv42, i32 5
%vecext44 = extractelement <16 x i8> %a, i32 6
%conv45 = zext i8 %vecext44 to i32
%vecext46 = extractelement <16 x i8> %b, i32 6
%conv47 = zext i8 %vecext46 to i32
%sub48 = sub nsw i32 %conv45, %conv47
%ispos181 = icmp sgt i32 %sub48, -1
%neg182 = sub nsw i32 0, %sub48
%6 = select i1 %ispos181, i32 %sub48, i32 %neg182
%conv50 = trunc i32 %6 to i8
%vecins51 = insertelement <16 x i8> %vecins43, i8 %conv50, i32 6
%vecext52 = extractelement <16 x i8> %a, i32 7
%conv53 = zext i8 %vecext52 to i32
%vecext54 = extractelement <16 x i8> %b, i32 7
%conv55 = zext i8 %vecext54 to i32
%sub56 = sub nsw i32 %conv53, %conv55
%ispos183 = icmp sgt i32 %sub56, -1
%neg184 = sub nsw i32 0, %sub56
%7 = select i1 %ispos183, i32 %sub56, i32 %neg184
%conv58 = trunc i32 %7 to i8
%vecins59 = insertelement <16 x i8> %vecins51, i8 %conv58, i32 7
%vecext60 = extractelement <16 x i8> %a, i32 8
%conv61 = zext i8 %vecext60 to i32
%vecext62 = extractelement <16 x i8> %b, i32 8
%conv63 = zext i8 %vecext62 to i32
%sub64 = sub nsw i32 %conv61, %conv63
%ispos185 = icmp sgt i32 %sub64, -1
%neg186 = sub nsw i32 0, %sub64
%8 = select i1 %ispos185, i32 %sub64, i32 %neg186
%conv66 = trunc i32 %8 to i8
%vecins67 = insertelement <16 x i8> %vecins59, i8 %conv66, i32 8
%vecext68 = extractelement <16 x i8> %a, i32 9
%conv69 = zext i8 %vecext68 to i32
%vecext70 = extractelement <16 x i8> %b, i32 9
%conv71 = zext i8 %vecext70 to i32
%sub72 = sub nsw i32 %conv69, %conv71
%ispos187 = icmp sgt i32 %sub72, -1
%neg188 = sub nsw i32 0, %sub72
%9 = select i1 %ispos187, i32 %sub72, i32 %neg188
%conv74 = trunc i32 %9 to i8
%vecins75 = insertelement <16 x i8> %vecins67, i8 %conv74, i32 9
%vecext76 = extractelement <16 x i8> %a, i32 10
%conv77 = zext i8 %vecext76 to i32
%vecext78 = extractelement <16 x i8> %b, i32 10
%conv79 = zext i8 %vecext78 to i32
%sub80 = sub nsw i32 %conv77, %conv79
%ispos189 = icmp sgt i32 %sub80, -1
%neg190 = sub nsw i32 0, %sub80
%10 = select i1 %ispos189, i32 %sub80, i32 %neg190
%conv82 = trunc i32 %10 to i8
%vecins83 = insertelement <16 x i8> %vecins75, i8 %conv82, i32 10
%vecext84 = extractelement <16 x i8> %a, i32 11
%conv85 = zext i8 %vecext84 to i32
%vecext86 = extractelement <16 x i8> %b, i32 11
%conv87 = zext i8 %vecext86 to i32
%sub88 = sub nsw i32 %conv85, %conv87
%ispos191 = icmp sgt i32 %sub88, -1
%neg192 = sub nsw i32 0, %sub88
%11 = select i1 %ispos191, i32 %sub88, i32 %neg192
%conv90 = trunc i32 %11 to i8
%vecins91 = insertelement <16 x i8> %vecins83, i8 %conv90, i32 11
%vecext92 = extractelement <16 x i8> %a, i32 12
%conv93 = zext i8 %vecext92 to i32
%vecext94 = extractelement <16 x i8> %b, i32 12
%conv95 = zext i8 %vecext94 to i32
%sub96 = sub nsw i32 %conv93, %conv95
%ispos193 = icmp sgt i32 %sub96, -1
%neg194 = sub nsw i32 0, %sub96
%12 = select i1 %ispos193, i32 %sub96, i32 %neg194
%conv98 = trunc i32 %12 to i8
%vecins99 = insertelement <16 x i8> %vecins91, i8 %conv98, i32 12
%vecext100 = extractelement <16 x i8> %a, i32 13
%conv101 = zext i8 %vecext100 to i32
%vecext102 = extractelement <16 x i8> %b, i32 13
%conv103 = zext i8 %vecext102 to i32
%sub104 = sub nsw i32 %conv101, %conv103
%ispos195 = icmp sgt i32 %sub104, -1
%neg196 = sub nsw i32 0, %sub104
%13 = select i1 %ispos195, i32 %sub104, i32 %neg196
%conv106 = trunc i32 %13 to i8
%vecins107 = insertelement <16 x i8> %vecins99, i8 %conv106, i32 13
%vecext108 = extractelement <16 x i8> %a, i32 14
%conv109 = zext i8 %vecext108 to i32
%vecext110 = extractelement <16 x i8> %b, i32 14
%conv111 = zext i8 %vecext110 to i32
%sub112 = sub nsw i32 %conv109, %conv111
%ispos197 = icmp sgt i32 %sub112, -1
%neg198 = sub nsw i32 0, %sub112
%14 = select i1 %ispos197, i32 %sub112, i32 %neg198
%conv114 = trunc i32 %14 to i8
%vecins115 = insertelement <16 x i8> %vecins107, i8 %conv114, i32 14
%vecext116 = extractelement <16 x i8> %a, i32 15
%conv117 = zext i8 %vecext116 to i32
%vecext118 = extractelement <16 x i8> %b, i32 15
%conv119 = zext i8 %vecext118 to i32
%sub120 = sub nsw i32 %conv117, %conv119
%ispos199 = icmp sgt i32 %sub120, -1
%neg200 = sub nsw i32 0, %sub120
%15 = select i1 %ispos199, i32 %sub120, i32 %neg200
%conv122 = trunc i32 %15 to i8
%vecins123 = insertelement <16 x i8> %vecins115, i8 %conv122, i32 15
ret <16 x i8> %vecins123
; CHECK-LABEL: sub_absv_8
; CHECK-NOT: vabsdub
; CHECK: subf
; CHECK-NOT: vabsdub
; CHECK: xor
; CHECK-NOT: vabsdub
; CHECK: blr
; CHECK-PWR8-LABEL: sub_absv_8
; CHECK-PWR8: subf
; CHECK-PWR8: xor
; CHECK-PWR8: blr
}
; Function Attrs: nounwind readnone
define <4 x i32> @sub_absv_vec_32(<4 x i32> %a, <4 x i32> %b) local_unnamed_addr {
entry:
%sub = sub <4 x i32> %a, %b
%sub.i = sub <4 x i32> zeroinitializer, %sub
%0 = tail call <4 x i32> @llvm.ppc.altivec.vmaxsw(<4 x i32> %sub, <4 x i32> %sub.i)
ret <4 x i32> %0
; CHECK-LABEL: sub_absv_vec_32
; CHECK: vabsduw 2, 2, 3
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: sub_absv_vec_32
; CHECK-PWR8: xxlxor
; CHECK-PWR8: vsubuwm
; CHECK-PWR8: vmaxsw
; CHECK-PWR8: blr
}
; Function Attrs: nounwind readnone
define <8 x i16> @sub_absv_vec_16(<8 x i16> %a, <8 x i16> %b) local_unnamed_addr {
entry:
%sub = sub <8 x i16> %a, %b
%sub.i = sub <8 x i16> zeroinitializer, %sub
%0 = tail call <8 x i16> @llvm.ppc.altivec.vmaxsh(<8 x i16> %sub, <8 x i16> %sub.i)
ret <8 x i16> %0
; CHECK-LABEL: sub_absv_vec_16
; CHECK: vabsduh 2, 2, 3
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: sub_absv_vec_16
; CHECK-PWR8: xxlxor
; CHECK-PWR8: vsubuhm
; CHECK-PWR8: vmaxsh
; CHECK-PWR8: blr
}
; Function Attrs: nounwind readnone
define <16 x i8> @sub_absv_vec_8(<16 x i8> %a, <16 x i8> %b) local_unnamed_addr {
entry:
%sub = sub <16 x i8> %a, %b
%sub.i = sub <16 x i8> zeroinitializer, %sub
%0 = tail call <16 x i8> @llvm.ppc.altivec.vmaxsb(<16 x i8> %sub, <16 x i8> %sub.i)
ret <16 x i8> %0
; CHECK-LABEL: sub_absv_vec_8
; CHECK: vabsdub 2, 2, 3
; CHECK-NEXT: blr
; CHECK-PWR8-LABEL: sub_absv_vec_8
; CHECK-PWR8: xxlxor
; CHECK-PWR8: vsububm
; CHECK-PWR8: vmaxsb
; CHECK-PWR8: blr
}
; Function Attrs: nounwind readnone
declare <4 x i32> @llvm.ppc.altivec.vmaxsw(<4 x i32>, <4 x i32>)
; Function Attrs: nounwind readnone
declare <8 x i16> @llvm.ppc.altivec.vmaxsh(<8 x i16>, <8 x i16>)
; Function Attrs: nounwind readnone
declare <16 x i8> @llvm.ppc.altivec.vmaxsb(<16 x i8>, <16 x i8>)
Reference in New Issue View Git Blame Copy Permalink