maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/bitcast-setcc-128.ll

870 lines
31 KiB
LLVM
Raw Normal View History

[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
[X86][SSE] Dropped -mcpu from bitcast+setcc tests Use triple and attribute only for consistency Added SSE2/AVX tests on 256-bit vectors to test PACKSS behaviour llvm-svn: 307289
2017-07-07 02:27:34 +08:00
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE2-SSSE3,SSE2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE2-SSSE3,SSSE3
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX12,AVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX12,AVX2
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512 --check-prefixes=AVX512F
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl,+avx512bw | FileCheck %s --check-prefixes=AVX512 --check-prefixes=AVX512BW
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
define i8 @v8i16(<8 x i16> %a, <8 x i16> %b) {
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; SSE2-SSSE3-LABEL: v8i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; SSE2-SSSE3-NEXT: pcmpgtw %xmm1, %xmm0
; SSE2-SSSE3-NEXT: packsswb %xmm0, %xmm0
; SSE2-SSSE3-NEXT: pmovmskb %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
;
; AVX12-LABEL: v8i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpcmpgtw %xmm1, %xmm0, %xmm0
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; AVX12-NEXT: vpacksswb %xmm0, %xmm0, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpmovmskb %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v8i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpcmpgtw %xmm1, %xmm0, %xmm0
[X86] When lowering truncates to vXi1, don't sign extend i16/i8 types to 512-bit if we have VLX. This should only affect what we do for v8i16. Previously we went to v8i64, but if we have VLX we only need v8i32. This prevents an unnecessary zmm usage. llvm-svn: 321303
2017-12-22 04:45:13 +08:00
; AVX512F-NEXT: vpmovsxwd %xmm0, %ymm0
; AVX512F-NEXT: vptestmd %ymm0, %ymm0, %k0
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v8i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpcmpgtw %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
%x = icmp sgt <8 x i16> %a, %b
%res = bitcast <8 x i1> %x to i8
ret i8 %res
}
define i4 @v4i32(<4 x i32> %a, <4 x i32> %b) {
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-LABEL: v4i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pcmpgtd %xmm1, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskps %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-LABEL: v4i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpcmpgtd %xmm1, %xmm0, %xmm0
; AVX12-NEXT: vmovmskps %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v4i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpcmpgtd %xmm1, %xmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v4i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpcmpgtd %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
%x = icmp sgt <4 x i32> %a, %b
%res = bitcast <4 x i1> %x to i4
ret i4 %res
}
define i4 @v4f32(<4 x float> %a, <4 x float> %b) {
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-LABEL: v4f32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: cmpltps %xmm0, %xmm1
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskps %xmm1, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-LABEL: v4f32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vcmpltps %xmm0, %xmm1, %xmm0
; AVX12-NEXT: vmovmskps %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v4f32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vcmpltps %xmm0, %xmm1, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v4f32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vcmpltps %xmm0, %xmm1, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
%x = fcmp ogt <4 x float> %a, %b
%res = bitcast <4 x i1> %x to i4
ret i4 %res
}
define i16 @v16i8(<16 x i8> %a, <16 x i8> %b) {
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-LABEL: v16i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pcmpgtb %xmm1, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: pmovmskb %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $ax killed $ax killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-LABEL: v16i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpcmpgtb %xmm1, %xmm0, %xmm0
; AVX12-NEXT: vpmovmskb %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $ax killed $ax killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v16i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpcmpgtb %xmm1, %xmm0, %xmm0
[X86] Allow combinevxi1Bitcast to use pmovmskb on avx512 targets if the input is a truncate from v16i8/v32i8. This is especially helpful on targets without avx512bw since we don't have a good way to convert from v16i8/v32i8 to v16i1/v32i1 for the truncate anyway. If we're just going to convert it to a GPR we might as well use pmovmskb to accomplish both. llvm-svn: 350480
2019-01-06 05:40:07 +08:00
; AVX512F-NEXT: vpmovmskb %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $ax killed $ax killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v16i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpcmpgtb %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $ax killed $ax killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
%x = icmp sgt <16 x i8> %a, %b
%res = bitcast <16 x i1> %x to i16
ret i16 %res
}
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
define i2 @v2i8(<2 x i8> %a, <2 x i8> %b) {
; SSE2-SSSE3-LABEL: v2i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: psllq $56, %xmm0
; SSE2-SSSE3-NEXT: movdqa %xmm0, %xmm2
; SSE2-SSSE3-NEXT: psrad $31, %xmm2
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,3,2,3]
; SSE2-SSSE3-NEXT: psrad $24, %xmm0
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,3]
; SSE2-SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
; SSE2-SSSE3-NEXT: psllq $56, %xmm1
; SSE2-SSSE3-NEXT: movdqa %xmm1, %xmm2
; SSE2-SSSE3-NEXT: psrad $31, %xmm2
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,3,2,3]
; SSE2-SSSE3-NEXT: psrad $24, %xmm1
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
; SSE2-SSSE3-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
[X86] When lowering unsigned v2i64 setcc without SSE42, flip the sign bits in the v2i64 type then bitcast to v4i32. This may give slightly better opportunities for DAG combine to simplify with the operations before the setcc. It also matches the type the xors will eventually be promoted to anyway so it saves a legalization step. Almost all of the test changes are because our constant pool entry is now v2i64 instead of v4i32 on 64-bit targets. On 32-bit targets getConstant should be emitting a v4i32 build_vector and a v4i32->v2i64 bitcast. There are a couple test cases where it appears we now combine a bitwise not with one of these xors which caused a new constant vector to be generated. This prevented a constant pool entry from being shared. But if that's an issue we're concerned about, it seems we need to address it another way that just relying a bitcast to hide it. This came about from experiments I've been trying with pushing the promotion of and/or/xor to vXi64 later than LegalizeVectorOps where it is today. We run LegalizeVectorOps in a bottom up order. So the and/or/xor are promoted before their users are legalized. The bitcasts added for the promotion act as a barrier to computeKnownBits if we try to use it during vector legalization of a later operation. So by moving the promotion out we can hopefully get better results from computeKnownBits/computeNumSignBits like in LowerTruncate on AVX512. I've also looked at running LegalizeVectorOps in a top down order like LegalizeDAG, but thats showing some other issues. llvm-svn: 344071
2018-10-10 03:05:50 +08:00
; SSE2-SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648]
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pxor %xmm2, %xmm1
; SSE2-SSSE3-NEXT: pxor %xmm2, %xmm0
; SSE2-SSSE3-NEXT: movdqa %xmm0, %xmm2
[TargetLowering] Add ISD::OR + ISD::XOR handling to SimplifyDemandedVectorElts Differential Revision: https://reviews.llvm.org/D55600 llvm-svn: 349264
2018-12-15 19:36:36 +08:00
; SSE2-SSSE3-NEXT: pcmpeqd %xmm1, %xmm2
; SSE2-SSSE3-NEXT: pcmpgtd %xmm1, %xmm0
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,2,2]
; SSE2-SSSE3-NEXT: pand %xmm2, %xmm1
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: por %xmm0, %xmm1
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskpd %xmm1, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
;
; AVX1-LABEL: v2i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX1: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-NEXT: vpsllq $56, %xmm1, %xmm1
; AVX1-NEXT: vpsrad $31, %xmm1, %xmm2
; AVX1-NEXT: vpsrad $24, %xmm1, %xmm1
; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
; AVX1-NEXT: vpsllq $56, %xmm0, %xmm0
; AVX1-NEXT: vpsrad $31, %xmm0, %xmm2
; AVX1-NEXT: vpsrad $24, %xmm0, %xmm0
; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm0, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX1-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX1-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-NEXT: retq
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-LABEL: v2i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX2: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-NEXT: vpsllq $56, %xmm1, %xmm1
; AVX2-NEXT: vpsrad $31, %xmm1, %xmm2
; AVX2-NEXT: vpsrad $24, %xmm1, %xmm1
; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
; AVX2-NEXT: vpsllq $56, %xmm0, %xmm0
; AVX2-NEXT: vpsrad $31, %xmm0, %xmm2
; AVX2-NEXT: vpsrad $24, %xmm0, %xmm0
; AVX2-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3]
; AVX2-NEXT: vpcmpgtq %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX2-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-NEXT: retq
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v2i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpsllq $56, %xmm1, %xmm1
; AVX512F-NEXT: vpsraq $56, %xmm1, %xmm1
; AVX512F-NEXT: vpsllq $56, %xmm0, %xmm0
; AVX512F-NEXT: vpsraq $56, %xmm0, %xmm0
; AVX512F-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v2i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpsllq $56, %xmm1, %xmm1
; AVX512BW-NEXT: vpsraq $56, %xmm1, %xmm1
; AVX512BW-NEXT: vpsllq $56, %xmm0, %xmm0
; AVX512BW-NEXT: vpsraq $56, %xmm0, %xmm0
; AVX512BW-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
%x = icmp sgt <2 x i8> %a, %b
%res = bitcast <2 x i1> %x to i2
ret i2 %res
}
define i2 @v2i16(<2 x i16> %a, <2 x i16> %b) {
; SSE2-SSSE3-LABEL: v2i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: psllq $48, %xmm0
; SSE2-SSSE3-NEXT: movdqa %xmm0, %xmm2
; SSE2-SSSE3-NEXT: psrad $31, %xmm2
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,3,2,3]
; SSE2-SSSE3-NEXT: psrad $16, %xmm0
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,3]
; SSE2-SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
; SSE2-SSSE3-NEXT: psllq $48, %xmm1
; SSE2-SSSE3-NEXT: movdqa %xmm1, %xmm2
; SSE2-SSSE3-NEXT: psrad $31, %xmm2
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,3,2,3]
; SSE2-SSSE3-NEXT: psrad $16, %xmm1
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
; SSE2-SSSE3-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
[X86] When lowering unsigned v2i64 setcc without SSE42, flip the sign bits in the v2i64 type then bitcast to v4i32. This may give slightly better opportunities for DAG combine to simplify with the operations before the setcc. It also matches the type the xors will eventually be promoted to anyway so it saves a legalization step. Almost all of the test changes are because our constant pool entry is now v2i64 instead of v4i32 on 64-bit targets. On 32-bit targets getConstant should be emitting a v4i32 build_vector and a v4i32->v2i64 bitcast. There are a couple test cases where it appears we now combine a bitwise not with one of these xors which caused a new constant vector to be generated. This prevented a constant pool entry from being shared. But if that's an issue we're concerned about, it seems we need to address it another way that just relying a bitcast to hide it. This came about from experiments I've been trying with pushing the promotion of and/or/xor to vXi64 later than LegalizeVectorOps where it is today. We run LegalizeVectorOps in a bottom up order. So the and/or/xor are promoted before their users are legalized. The bitcasts added for the promotion act as a barrier to computeKnownBits if we try to use it during vector legalization of a later operation. So by moving the promotion out we can hopefully get better results from computeKnownBits/computeNumSignBits like in LowerTruncate on AVX512. I've also looked at running LegalizeVectorOps in a top down order like LegalizeDAG, but thats showing some other issues. llvm-svn: 344071
2018-10-10 03:05:50 +08:00
; SSE2-SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648]
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pxor %xmm2, %xmm1
; SSE2-SSSE3-NEXT: pxor %xmm2, %xmm0
; SSE2-SSSE3-NEXT: movdqa %xmm0, %xmm2
[TargetLowering] Add ISD::OR + ISD::XOR handling to SimplifyDemandedVectorElts Differential Revision: https://reviews.llvm.org/D55600 llvm-svn: 349264
2018-12-15 19:36:36 +08:00
; SSE2-SSSE3-NEXT: pcmpeqd %xmm1, %xmm2
; SSE2-SSSE3-NEXT: pcmpgtd %xmm1, %xmm0
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,2,2]
; SSE2-SSSE3-NEXT: pand %xmm2, %xmm1
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: por %xmm0, %xmm1
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskpd %xmm1, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
;
; AVX1-LABEL: v2i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX1: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-NEXT: vpsllq $48, %xmm1, %xmm1
; AVX1-NEXT: vpsrad $31, %xmm1, %xmm2
; AVX1-NEXT: vpsrad $16, %xmm1, %xmm1
; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
; AVX1-NEXT: vpsllq $48, %xmm0, %xmm0
; AVX1-NEXT: vpsrad $31, %xmm0, %xmm2
; AVX1-NEXT: vpsrad $16, %xmm0, %xmm0
; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm0, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX1-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX1-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-NEXT: retq
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-LABEL: v2i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX2: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-NEXT: vpsllq $48, %xmm1, %xmm1
; AVX2-NEXT: vpsrad $31, %xmm1, %xmm2
; AVX2-NEXT: vpsrad $16, %xmm1, %xmm1
; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
; AVX2-NEXT: vpsllq $48, %xmm0, %xmm0
; AVX2-NEXT: vpsrad $31, %xmm0, %xmm2
; AVX2-NEXT: vpsrad $16, %xmm0, %xmm0
; AVX2-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3]
; AVX2-NEXT: vpcmpgtq %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX2-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-NEXT: retq
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v2i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpsllq $48, %xmm1, %xmm1
; AVX512F-NEXT: vpsraq $48, %xmm1, %xmm1
; AVX512F-NEXT: vpsllq $48, %xmm0, %xmm0
; AVX512F-NEXT: vpsraq $48, %xmm0, %xmm0
; AVX512F-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v2i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpsllq $48, %xmm1, %xmm1
; AVX512BW-NEXT: vpsraq $48, %xmm1, %xmm1
; AVX512BW-NEXT: vpsllq $48, %xmm0, %xmm0
; AVX512BW-NEXT: vpsraq $48, %xmm0, %xmm0
; AVX512BW-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
%x = icmp sgt <2 x i16> %a, %b
%res = bitcast <2 x i1> %x to i2
ret i2 %res
}
define i2 @v2i32(<2 x i32> %a, <2 x i32> %b) {
; SSE2-SSSE3-LABEL: v2i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: psllq $32, %xmm0
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,3,2,3]
; SSE2-SSSE3-NEXT: psrad $31, %xmm0
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,3]
; SSE2-SSSE3-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1]
; SSE2-SSSE3-NEXT: psllq $32, %xmm1
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,3,2,3]
; SSE2-SSSE3-NEXT: psrad $31, %xmm1
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
; SSE2-SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
[X86] When lowering unsigned v2i64 setcc without SSE42, flip the sign bits in the v2i64 type then bitcast to v4i32. This may give slightly better opportunities for DAG combine to simplify with the operations before the setcc. It also matches the type the xors will eventually be promoted to anyway so it saves a legalization step. Almost all of the test changes are because our constant pool entry is now v2i64 instead of v4i32 on 64-bit targets. On 32-bit targets getConstant should be emitting a v4i32 build_vector and a v4i32->v2i64 bitcast. There are a couple test cases where it appears we now combine a bitwise not with one of these xors which caused a new constant vector to be generated. This prevented a constant pool entry from being shared. But if that's an issue we're concerned about, it seems we need to address it another way that just relying a bitcast to hide it. This came about from experiments I've been trying with pushing the promotion of and/or/xor to vXi64 later than LegalizeVectorOps where it is today. We run LegalizeVectorOps in a bottom up order. So the and/or/xor are promoted before their users are legalized. The bitcasts added for the promotion act as a barrier to computeKnownBits if we try to use it during vector legalization of a later operation. So by moving the promotion out we can hopefully get better results from computeKnownBits/computeNumSignBits like in LowerTruncate on AVX512. I've also looked at running LegalizeVectorOps in a top down order like LegalizeDAG, but thats showing some other issues. llvm-svn: 344071
2018-10-10 03:05:50 +08:00
; SSE2-SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [2147483648,2147483648]
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pxor %xmm1, %xmm0
; SSE2-SSSE3-NEXT: pxor %xmm1, %xmm2
; SSE2-SSSE3-NEXT: movdqa %xmm2, %xmm1
[TargetLowering] Add ISD::OR + ISD::XOR handling to SimplifyDemandedVectorElts Differential Revision: https://reviews.llvm.org/D55600 llvm-svn: 349264
2018-12-15 19:36:36 +08:00
; SSE2-SSSE3-NEXT: pcmpeqd %xmm0, %xmm1
; SSE2-SSSE3-NEXT: pcmpgtd %xmm0, %xmm2
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,0,2,2]
; SSE2-SSSE3-NEXT: pand %xmm1, %xmm0
; SSE2-SSSE3-NEXT: por %xmm2, %xmm0
; SSE2-SSSE3-NEXT: movmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-LABEL: v2i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX1: # %bb.0:
[X86][SSE] Combine (some) target shuffles with multiple uses As discussed on D41794, we have many cases where we fail to combine shuffles as the input operands have other uses. This patch permits these shuffles to be combined as long as they don't introduce additional variable shuffle masks, which should reduce instruction dependencies and allow the total number of shuffles to still drop without increasing the constant pool. However, this may mean that some memory folds may no longer occur, and on pre-AVX require the occasional extra register move. This also exposes some poor PMULDQ/PMULUDQ codegen which was doing unnecessary upper/lower calculations which will in fact fold to zero/undef - the fix will be added in a followup commit. Differential Revision: https://reviews.llvm.org/D50328 llvm-svn: 339335
2018-08-09 20:30:02 +08:00
; AVX1-NEXT: vpsllq $32, %xmm1, %xmm2
; AVX1-NEXT: vpsrad $31, %xmm2, %xmm2
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
[X86][SSE] Combine (some) target shuffles with multiple uses As discussed on D41794, we have many cases where we fail to combine shuffles as the input operands have other uses. This patch permits these shuffles to be combined as long as they don't introduce additional variable shuffle masks, which should reduce instruction dependencies and allow the total number of shuffles to still drop without increasing the constant pool. However, this may mean that some memory folds may no longer occur, and on pre-AVX require the occasional extra register move. This also exposes some poor PMULDQ/PMULUDQ codegen which was doing unnecessary upper/lower calculations which will in fact fold to zero/undef - the fix will be added in a followup commit. Differential Revision: https://reviews.llvm.org/D50328 llvm-svn: 339335
2018-08-09 20:30:02 +08:00
; AVX1-NEXT: vpsllq $32, %xmm0, %xmm2
; AVX1-NEXT: vpsrad $31, %xmm2, %xmm2
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm0, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX1-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX1-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; AVX1-NEXT: retq
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-LABEL: v2i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX2: # %bb.0:
[X86][SSE] Combine (some) target shuffles with multiple uses As discussed on D41794, we have many cases where we fail to combine shuffles as the input operands have other uses. This patch permits these shuffles to be combined as long as they don't introduce additional variable shuffle masks, which should reduce instruction dependencies and allow the total number of shuffles to still drop without increasing the constant pool. However, this may mean that some memory folds may no longer occur, and on pre-AVX require the occasional extra register move. This also exposes some poor PMULDQ/PMULUDQ codegen which was doing unnecessary upper/lower calculations which will in fact fold to zero/undef - the fix will be added in a followup commit. Differential Revision: https://reviews.llvm.org/D50328 llvm-svn: 339335
2018-08-09 20:30:02 +08:00
; AVX2-NEXT: vpsllq $32, %xmm1, %xmm2
; AVX2-NEXT: vpsrad $31, %xmm2, %xmm2
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
[X86][SSE] Combine (some) target shuffles with multiple uses As discussed on D41794, we have many cases where we fail to combine shuffles as the input operands have other uses. This patch permits these shuffles to be combined as long as they don't introduce additional variable shuffle masks, which should reduce instruction dependencies and allow the total number of shuffles to still drop without increasing the constant pool. However, this may mean that some memory folds may no longer occur, and on pre-AVX require the occasional extra register move. This also exposes some poor PMULDQ/PMULUDQ codegen which was doing unnecessary upper/lower calculations which will in fact fold to zero/undef - the fix will be added in a followup commit. Differential Revision: https://reviews.llvm.org/D50328 llvm-svn: 339335
2018-08-09 20:30:02 +08:00
; AVX2-NEXT: vpsllq $32, %xmm0, %xmm2
; AVX2-NEXT: vpsrad $31, %xmm2, %xmm2
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3]
; AVX2-NEXT: vpcmpgtq %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX2-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX2-NEXT: retq
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v2i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpsllq $32, %xmm1, %xmm1
; AVX512F-NEXT: vpsraq $32, %xmm1, %xmm1
; AVX512F-NEXT: vpsllq $32, %xmm0, %xmm0
; AVX512F-NEXT: vpsraq $32, %xmm0, %xmm0
; AVX512F-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v2i32:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpsllq $32, %xmm1, %xmm1
; AVX512BW-NEXT: vpsraq $32, %xmm1, %xmm1
; AVX512BW-NEXT: vpsllq $32, %xmm0, %xmm0
; AVX512BW-NEXT: vpsraq $32, %xmm0, %xmm0
; AVX512BW-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
%x = icmp sgt <2 x i32> %a, %b
%res = bitcast <2 x i1> %x to i2
ret i2 %res
}
define i2 @v2i64(<2 x i64> %a, <2 x i64> %b) {
; SSE2-SSSE3-LABEL: v2i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] When lowering unsigned v2i64 setcc without SSE42, flip the sign bits in the v2i64 type then bitcast to v4i32. This may give slightly better opportunities for DAG combine to simplify with the operations before the setcc. It also matches the type the xors will eventually be promoted to anyway so it saves a legalization step. Almost all of the test changes are because our constant pool entry is now v2i64 instead of v4i32 on 64-bit targets. On 32-bit targets getConstant should be emitting a v4i32 build_vector and a v4i32->v2i64 bitcast. There are a couple test cases where it appears we now combine a bitwise not with one of these xors which caused a new constant vector to be generated. This prevented a constant pool entry from being shared. But if that's an issue we're concerned about, it seems we need to address it another way that just relying a bitcast to hide it. This came about from experiments I've been trying with pushing the promotion of and/or/xor to vXi64 later than LegalizeVectorOps where it is today. We run LegalizeVectorOps in a bottom up order. So the and/or/xor are promoted before their users are legalized. The bitcasts added for the promotion act as a barrier to computeKnownBits if we try to use it during vector legalization of a later operation. So by moving the promotion out we can hopefully get better results from computeKnownBits/computeNumSignBits like in LowerTruncate on AVX512. I've also looked at running LegalizeVectorOps in a top down order like LegalizeDAG, but thats showing some other issues. llvm-svn: 344071
2018-10-10 03:05:50 +08:00
; SSE2-SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648]
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pxor %xmm2, %xmm1
; SSE2-SSSE3-NEXT: pxor %xmm2, %xmm0
; SSE2-SSSE3-NEXT: movdqa %xmm0, %xmm2
[TargetLowering] Add ISD::OR + ISD::XOR handling to SimplifyDemandedVectorElts Differential Revision: https://reviews.llvm.org/D55600 llvm-svn: 349264
2018-12-15 19:36:36 +08:00
; SSE2-SSSE3-NEXT: pcmpeqd %xmm1, %xmm2
; SSE2-SSSE3-NEXT: pcmpgtd %xmm1, %xmm0
; SSE2-SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,0,2,2]
; SSE2-SSSE3-NEXT: pand %xmm2, %xmm1
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: por %xmm0, %xmm1
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskpd %xmm1, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-LABEL: v2i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpcmpgtq %xmm1, %xmm0, %xmm0
; AVX12-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v2i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v2i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpcmpgtq %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
%x = icmp sgt <2 x i64> %a, %b
%res = bitcast <2 x i1> %x to i2
ret i2 %res
}
define i2 @v2f64(<2 x double> %a, <2 x double> %b) {
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-LABEL: v2f64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: cmpltpd %xmm0, %xmm1
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskpd %xmm1, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-LABEL: v2f64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vcmpltpd %xmm0, %xmm1, %xmm0
; AVX12-NEXT: vmovmskpd %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v2f64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vcmpltpd %xmm0, %xmm1, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v2f64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vcmpltpd %xmm0, %xmm1, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Adding tests for scalar bitcasts from vsetcc. NFC. llvm-svn: 303328
2017-05-18 15:04:48 +08:00
%x = fcmp ogt <2 x double> %a, %b
%res = bitcast <2 x i1> %x to i2
ret i2 %res
}
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
define i4 @v4i8(<4 x i8> %a, <4 x i8> %b) {
; SSE2-SSSE3-LABEL: v4i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pslld $24, %xmm1
; SSE2-SSSE3-NEXT: psrad $24, %xmm1
; SSE2-SSSE3-NEXT: pslld $24, %xmm0
; SSE2-SSSE3-NEXT: psrad $24, %xmm0
; SSE2-SSSE3-NEXT: pcmpgtd %xmm1, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskps %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-LABEL: v4i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpslld $24, %xmm1, %xmm1
; AVX12-NEXT: vpsrad $24, %xmm1, %xmm1
; AVX12-NEXT: vpslld $24, %xmm0, %xmm0
; AVX12-NEXT: vpsrad $24, %xmm0, %xmm0
; AVX12-NEXT: vpcmpgtd %xmm1, %xmm0, %xmm0
; AVX12-NEXT: vmovmskps %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v4i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpslld $24, %xmm1, %xmm1
; AVX512F-NEXT: vpsrad $24, %xmm1, %xmm1
; AVX512F-NEXT: vpslld $24, %xmm0, %xmm0
; AVX512F-NEXT: vpsrad $24, %xmm0, %xmm0
; AVX512F-NEXT: vpcmpgtd %xmm1, %xmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v4i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpslld $24, %xmm1, %xmm1
; AVX512BW-NEXT: vpsrad $24, %xmm1, %xmm1
; AVX512BW-NEXT: vpslld $24, %xmm0, %xmm0
; AVX512BW-NEXT: vpsrad $24, %xmm0, %xmm0
; AVX512BW-NEXT: vpcmpgtd %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
%x = icmp sgt <4 x i8> %a, %b
%res = bitcast <4 x i1> %x to i4
ret i4 %res
}
define i4 @v4i16(<4 x i16> %a, <4 x i16> %b) {
; SSE2-SSSE3-LABEL: v4i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: pslld $16, %xmm1
; SSE2-SSSE3-NEXT: psrad $16, %xmm1
; SSE2-SSSE3-NEXT: pslld $16, %xmm0
; SSE2-SSSE3-NEXT: psrad $16, %xmm0
; SSE2-SSSE3-NEXT: pcmpgtd %xmm1, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; SSE2-SSSE3-NEXT: movmskps %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
; SSE2-SSSE3-NEXT: retq
;
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-LABEL: v4i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpslld $16, %xmm1, %xmm1
; AVX12-NEXT: vpsrad $16, %xmm1, %xmm1
; AVX12-NEXT: vpslld $16, %xmm0, %xmm0
; AVX12-NEXT: vpsrad $16, %xmm0, %xmm0
; AVX12-NEXT: vpcmpgtd %xmm1, %xmm0, %xmm0
; AVX12-NEXT: vmovmskps %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v4i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpslld $16, %xmm1, %xmm1
; AVX512F-NEXT: vpsrad $16, %xmm1, %xmm1
; AVX512F-NEXT: vpslld $16, %xmm0, %xmm0
; AVX512F-NEXT: vpsrad $16, %xmm0, %xmm0
; AVX512F-NEXT: vpcmpgtd %xmm1, %xmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v4i16:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpslld $16, %xmm1, %xmm1
; AVX512BW-NEXT: vpsrad $16, %xmm1, %xmm1
; AVX512BW-NEXT: vpslld $16, %xmm0, %xmm0
; AVX512BW-NEXT: vpsrad $16, %xmm0, %xmm0
; AVX512BW-NEXT: vpcmpgtd %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
%x = icmp sgt <4 x i16> %a, %b
%res = bitcast <4 x i1> %x to i4
ret i4 %res
}
define i8 @v8i8(<8 x i8> %a, <8 x i8> %b) {
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; SSE2-SSSE3-LABEL: v8i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; SSE2-SSSE3: # %bb.0:
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; SSE2-SSSE3-NEXT: psllw $8, %xmm1
; SSE2-SSSE3-NEXT: psraw $8, %xmm1
; SSE2-SSSE3-NEXT: psllw $8, %xmm0
; SSE2-SSSE3-NEXT: psraw $8, %xmm0
; SSE2-SSSE3-NEXT: pcmpgtw %xmm1, %xmm0
; SSE2-SSSE3-NEXT: packsswb %xmm0, %xmm0
; SSE2-SSSE3-NEXT: pmovmskb %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; SSE2-SSSE3-NEXT: # kill: def $al killed $al killed $eax
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; SSE2-SSSE3-NEXT: retq
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
;
; AVX12-LABEL: v8i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX12: # %bb.0:
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpsllw $8, %xmm1, %xmm1
; AVX12-NEXT: vpsraw $8, %xmm1, %xmm1
; AVX12-NEXT: vpsllw $8, %xmm0, %xmm0
; AVX12-NEXT: vpsraw $8, %xmm0, %xmm0
; AVX12-NEXT: vpcmpgtw %xmm1, %xmm0, %xmm0
[X86][SSE] Add support for lowering shuffles to PACKSS/PACKUS If the upper bits of a truncation shuffle patterns have at least the minimum number of sign/zero bits on their inputs then we can safely use PACKSS/PACKUS as shuffles. Partial fix for https://bugs.llvm.org/show_bug.cgi?id=34773 Differential Revision: https://reviews.llvm.org/D38472 llvm-svn: 314788
2017-10-03 20:01:31 +08:00
; AVX12-NEXT: vpacksswb %xmm0, %xmm0, %xmm0
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: vpmovmskb %xmm0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX12-NEXT: # kill: def $al killed $al killed $eax
[X86] Match bitcast of vxi1 to pmovmsk Summary: Add an early combine to match patterns such as: (i16 bitcast (v16i1 x)) -> (i16 movmsk (v16i8 sext (v16i1 x))) This combine needs to happen early enough before type-legalization scalarizes the result of the setcc. Reviewers: igorb, craig.topper, RKSimon Subscribers: delena, llvm-commits Differential Revision: https://reviews.llvm.org/D33311 llvm-svn: 304406
2017-06-01 19:27:57 +08:00
; AVX12-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
;
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-LABEL: v8i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512F: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vpsllw $8, %xmm1, %xmm1
; AVX512F-NEXT: vpsraw $8, %xmm1, %xmm1
; AVX512F-NEXT: vpsllw $8, %xmm0, %xmm0
; AVX512F-NEXT: vpsraw $8, %xmm0, %xmm0
; AVX512F-NEXT: vpcmpgtw %xmm1, %xmm0, %xmm0
[X86] When lowering truncates to vXi1, don't sign extend i16/i8 types to 512-bit if we have VLX. This should only affect what we do for v8i16. Previously we went to v8i64, but if we have VLX we only need v8i32. This prevents an unnecessary zmm usage. llvm-svn: 321303
2017-12-22 04:45:13 +08:00
; AVX512F-NEXT: vpmovsxwd %xmm0, %ymm0
; AVX512F-NEXT: vptestmd %ymm0, %ymm0, %k0
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: kmovw %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512F-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v8i8:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; AVX512BW: # %bb.0:
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: vpsllw $8, %xmm1, %xmm1
; AVX512BW-NEXT: vpsraw $8, %xmm1, %xmm1
; AVX512BW-NEXT: vpsllw $8, %xmm0, %xmm0
; AVX512BW-NEXT: vpsraw $8, %xmm0, %xmm0
; AVX512BW-NEXT: vpcmpgtw %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
Followup on Proposal to move MIR physical register namespace to '$' sigil. Discussed here: http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html In preparation for adding support for named vregs we are changing the sigil for physical registers in MIR to '$' from '%'. This will prevent name clashes of named physical register with named vregs. llvm-svn: 323922
2018-02-01 06:04:26 +08:00
; AVX512BW-NEXT: # kill: def $al killed $al killed $eax
[X86][AVX512] Split AVX512F and AVX512BW bool-vector bitcast tests llvm-svn: 317020
2017-11-01 02:41:48 +08:00
; AVX512BW-NEXT: retq
[X86] Add (ix bitcast(vsetcc)) test cases with illegal types. NFC. llvm-svn: 303530
2017-05-22 14:39:12 +08:00
%x = icmp sgt <8 x i8> %a, %b
%res = bitcast <8 x i1> %x to i8
ret i8 %res
}
[X86] Remove VT.isSimple() check from detectAVGPattern. Which types are considered 'simple' is a function of the requirements of all targets that LLVM supports. That shouldn't directly affect what types we are able to handle. The remainder of this code checks that the number of elements is a power of 2 and takes care of splitting down to a legal size. llvm-svn: 326063
2018-02-26 10:16:31 +08:00
define i64 @v16i8_widened_with_zeroes(<16 x i8> %a, <16 x i8> %b) {
; SSE2-SSSE3-LABEL: v16i8_widened_with_zeroes:
; SSE2-SSSE3: # %bb.0: # %entry
; SSE2-SSSE3-NEXT: pcmpeqb %xmm1, %xmm0
[X86] Add constant folding to combineMOVMSK. There's still some shortcoming in our ability to combine binops of constants with different sizes separated by an extend. I'll try to look at that next. llvm-svn: 326128
2018-02-27 05:17:33 +08:00
; SSE2-SSSE3-NEXT: pmovmskb %xmm0, %eax
[X86] Remove VT.isSimple() check from detectAVGPattern. Which types are considered 'simple' is a function of the requirements of all targets that LLVM supports. That shouldn't directly affect what types we are able to handle. The remainder of this code checks that the number of elements is a power of 2 and takes care of splitting down to a legal size. llvm-svn: 326063
2018-02-26 10:16:31 +08:00
; SSE2-SSSE3-NEXT: retq
;
[X86] Add a custom legalization for (i16 (bitcast v16i1)) and (i32 (bitcast v32i1)) without AVX512 to prevent scalarization Summary: We have an early DAG combine to turn these patterns into MOVMSK, but that combine doesn't work if the vXi1 type has more elements than the widest legal vXi8 type. Type legalization will eventually split it down to v16i1 or v32i1 and then the bitcast gets legalized to a truncstore and a scalar load. The truncstore will get lowered to a series of extracts and bit math. This patch adds a custom legalization to use a sign extend and MOVMSK instead. This prevents the eventual scalarization. Reviewers: spatel, RKSimon, zvi Reviewed By: RKSimon Subscribers: mgorny, llvm-commits Differential Revision: https://reviews.llvm.org/D43593 llvm-svn: 326119
2018-02-27 04:32:27 +08:00
; AVX1-LABEL: v16i8_widened_with_zeroes:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vpcmpeqb %xmm1, %xmm0, %xmm0
[X86] Add constant folding to combineMOVMSK. There's still some shortcoming in our ability to combine binops of constants with different sizes separated by an extend. I'll try to look at that next. llvm-svn: 326128
2018-02-27 05:17:33 +08:00
; AVX1-NEXT: vpmovmskb %xmm0, %eax
[X86] Add a custom legalization for (i16 (bitcast v16i1)) and (i32 (bitcast v32i1)) without AVX512 to prevent scalarization Summary: We have an early DAG combine to turn these patterns into MOVMSK, but that combine doesn't work if the vXi1 type has more elements than the widest legal vXi8 type. Type legalization will eventually split it down to v16i1 or v32i1 and then the bitcast gets legalized to a truncstore and a scalar load. The truncstore will get lowered to a series of extracts and bit math. This patch adds a custom legalization to use a sign extend and MOVMSK instead. This prevents the eventual scalarization. Reviewers: spatel, RKSimon, zvi Reviewed By: RKSimon Subscribers: mgorny, llvm-commits Differential Revision: https://reviews.llvm.org/D43593 llvm-svn: 326119
2018-02-27 04:32:27 +08:00
; AVX1-NEXT: retq
;
; AVX2-LABEL: v16i8_widened_with_zeroes:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpcmpeqb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vpmovmskb %ymm0, %eax
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
[X86] Remove VT.isSimple() check from detectAVGPattern. Which types are considered 'simple' is a function of the requirements of all targets that LLVM supports. That shouldn't directly affect what types we are able to handle. The remainder of this code checks that the number of elements is a power of 2 and takes care of splitting down to a legal size. llvm-svn: 326063
2018-02-26 10:16:31 +08:00
;
; AVX512F-LABEL: v16i8_widened_with_zeroes:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vpcmpeqb %xmm1, %xmm0, %xmm0
; AVX512F-NEXT: vpmovsxbd %xmm0, %zmm0
; AVX512F-NEXT: vptestmd %zmm0, %zmm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
; AVX512F-NEXT: movzwl %ax, %eax
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v16i8_widened_with_zeroes:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vpcmpeqb %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kmovq %k0, %rax
; AVX512BW-NEXT: retq
entry:
%c = icmp eq <16 x i8> %a, %b
%d = shufflevector <16 x i1> %c, <16 x i1> zeroinitializer, <64 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
%e = bitcast <64 x i1> %d to i64
ret i64 %e
}
[X86] Add constant folding to combineMOVMSK. There's still some shortcoming in our ability to combine binops of constants with different sizes separated by an extend. I'll try to look at that next. llvm-svn: 326128
2018-02-27 05:17:33 +08:00
define i64 @v16i8_widened_with_ones(<16 x i8> %a, <16 x i8> %b) {
; SSE2-SSSE3-LABEL: v16i8_widened_with_ones:
; SSE2-SSSE3: # %bb.0: # %entry
; SSE2-SSSE3-NEXT: pcmpeqb %xmm1, %xmm0
; SSE2-SSSE3-NEXT: pmovmskb %xmm0, %ecx
; SSE2-SSSE3-NEXT: orl $-65536, %ecx # imm = 0xFFFF0000
; SSE2-SSSE3-NEXT: movabsq $-4294967296, %rax # imm = 0xFFFFFFFF00000000
; SSE2-SSSE3-NEXT: orq %rcx, %rax
; SSE2-SSSE3-NEXT: retq
;
; AVX1-LABEL: v16i8_widened_with_ones:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vpcmpeqb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vpsllw $7, %xmm0, %xmm0
; AVX1-NEXT: vpmovmskb %xmm0, %ecx
; AVX1-NEXT: orl $-65536, %ecx # imm = 0xFFFF0000
; AVX1-NEXT: movabsq $-4294967296, %rax # imm = 0xFFFFFFFF00000000
; AVX1-NEXT: orq %rcx, %rax
; AVX1-NEXT: retq
;
; AVX2-LABEL: v16i8_widened_with_ones:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpcmpeqb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vinserti128 $1, {{.*}}(%rip), %ymm0, %ymm0
; AVX2-NEXT: vpsllw $7, %ymm0, %ymm0
; AVX2-NEXT: vpmovmskb %ymm0, %ecx
; AVX2-NEXT: movabsq $-4294967296, %rax # imm = 0xFFFFFFFF00000000
; AVX2-NEXT: orq %rcx, %rax
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: v16i8_widened_with_ones:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vpcmpeqb %xmm1, %xmm0, %xmm0
; AVX512F-NEXT: vpmovsxbd %xmm0, %zmm0
; AVX512F-NEXT: vptestmd %zmm0, %zmm0, %k0
; AVX512F-NEXT: kmovw %k0, %ecx
; AVX512F-NEXT: orl $-65536, %ecx # imm = 0xFFFF0000
; AVX512F-NEXT: movabsq $-4294967296, %rax # imm = 0xFFFFFFFF00000000
; AVX512F-NEXT: orq %rcx, %rax
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: v16i8_widened_with_ones:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vpcmpeqb %xmm1, %xmm0, %k0
; AVX512BW-NEXT: kxnorw %k0, %k0, %k1
; AVX512BW-NEXT: kunpckwd %k0, %k1, %k0
; AVX512BW-NEXT: kxnord %k0, %k0, %k1
; AVX512BW-NEXT: kunpckdq %k0, %k1, %k0
; AVX512BW-NEXT: kmovq %k0, %rax
; AVX512BW-NEXT: retq
entry:
%c = icmp eq <16 x i8> %a, %b
%d = shufflevector <16 x i1> %c, <16 x i1> <i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1, i1 -1>, <64 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
%e = bitcast <64 x i1> %d to i64
ret i64 %e
}
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
define void @bitcast_16i8_store(i16* %p, <16 x i8> %a0) {
; SSE2-SSSE3-LABEL: bitcast_16i8_store:
; SSE2-SSSE3: # %bb.0:
[X86] Prevent DAG combine from folding a bitcast from vXi1 to iX with a store on pre-AVX512 targets. If we fold the bitcast into the store we'll end up creating a truncating store to vXi1 that will get scalarized. Instead allow the bitcast to be turned into a movmsk. We probably need to do something if the store itself is a vXi1 type, but I'll leave that til a testcase appears. llvm-svn: 347632
2018-11-27 10:57:27 +08:00
; SSE2-SSSE3-NEXT: pmovmskb %xmm0, %eax
; SSE2-SSSE3-NEXT: movw %ax, (%rdi)
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
; SSE2-SSSE3-NEXT: retq
;
; AVX12-LABEL: bitcast_16i8_store:
; AVX12: # %bb.0:
[X86] Prevent DAG combine from folding a bitcast from vXi1 to iX with a store on pre-AVX512 targets. If we fold the bitcast into the store we'll end up creating a truncating store to vXi1 that will get scalarized. Instead allow the bitcast to be turned into a movmsk. We probably need to do something if the store itself is a vXi1 type, but I'll leave that til a testcase appears. llvm-svn: 347632
2018-11-27 10:57:27 +08:00
; AVX12-NEXT: vpmovmskb %xmm0, %eax
; AVX12-NEXT: movw %ax, (%rdi)
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
; AVX12-NEXT: retq
;
; AVX512F-LABEL: bitcast_16i8_store:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512F-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm0
; AVX512F-NEXT: vpmovsxbd %xmm0, %zmm0
; AVX512F-NEXT: vptestmd %zmm0, %zmm0, %k0
; AVX512F-NEXT: kmovw %k0, (%rdi)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: bitcast_16i8_store:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vpmovb2m %xmm0, %k0
; AVX512BW-NEXT: kmovw %k0, (%rdi)
; AVX512BW-NEXT: retq
%a1 = icmp slt <16 x i8> %a0, zeroinitializer
%a2 = bitcast <16 x i1> %a1 to i16
store i16 %a2, i16* %p
ret void
}
define void @bitcast_8i16_store(i8* %p, <8 x i16> %a0) {
; SSE2-SSSE3-LABEL: bitcast_8i16_store:
; SSE2-SSSE3: # %bb.0:
[X86][SSE] SimplifyDemandedBitsForTargetNode - Add initial PACKSS support In the case where we only want the sign bit (e.g. when using PACKSS truncation of comparison results for MOVMSK) then we can just demand the sign bit of the source operands. This makes use of the fact that PACKSS saturates out of range values to the min/max int values - so the sign bit is always preserved. Differential Revision: https://reviews.llvm.org/D60333 llvm-svn: 357859
2019-04-07 18:40:01 +08:00
; SSE2-SSSE3-NEXT: packsswb %xmm0, %xmm0
; SSE2-SSSE3-NEXT: pmovmskb %xmm0, %eax
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
; SSE2-SSSE3-NEXT: movb %al, (%rdi)
; SSE2-SSSE3-NEXT: retq
;
; AVX12-LABEL: bitcast_8i16_store:
; AVX12: # %bb.0:
; AVX12-NEXT: vpacksswb %xmm0, %xmm0, %xmm0
; AVX12-NEXT: vpmovmskb %xmm0, %eax
; AVX12-NEXT: movb %al, (%rdi)
; AVX12-NEXT: retq
;
; AVX512F-LABEL: bitcast_8i16_store:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512F-NEXT: vpcmpgtw %xmm0, %xmm1, %xmm0
; AVX512F-NEXT: vpmovsxwd %xmm0, %ymm0
; AVX512F-NEXT: vptestmd %ymm0, %ymm0, %k0
; AVX512F-NEXT: kmovw %k0, %eax
; AVX512F-NEXT: movb %al, (%rdi)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: bitcast_8i16_store:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vpmovw2m %xmm0, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
; AVX512BW-NEXT: movb %al, (%rdi)
; AVX512BW-NEXT: retq
%a1 = icmp slt <8 x i16> %a0, zeroinitializer
%a2 = bitcast <8 x i1> %a1 to i8
store i8 %a2, i8* %p
ret void
}
define void @bitcast_4i32_store(i4* %p, <4 x i32> %a0) {
; SSE2-SSSE3-LABEL: bitcast_4i32_store:
; SSE2-SSSE3: # %bb.0:
[X86] Prevent DAG combine from folding a bitcast from vXi1 to iX with a store on pre-AVX512 targets. If we fold the bitcast into the store we'll end up creating a truncating store to vXi1 that will get scalarized. Instead allow the bitcast to be turned into a movmsk. We probably need to do something if the store itself is a vXi1 type, but I'll leave that til a testcase appears. llvm-svn: 347632
2018-11-27 10:57:27 +08:00
; SSE2-SSSE3-NEXT: movmskps %xmm0, %eax
; SSE2-SSSE3-NEXT: movb %al, (%rdi)
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
; SSE2-SSSE3-NEXT: retq
;
; AVX12-LABEL: bitcast_4i32_store:
; AVX12: # %bb.0:
[X86] Prevent DAG combine from folding a bitcast from vXi1 to iX with a store on pre-AVX512 targets. If we fold the bitcast into the store we'll end up creating a truncating store to vXi1 that will get scalarized. Instead allow the bitcast to be turned into a movmsk. We probably need to do something if the store itself is a vXi1 type, but I'll leave that til a testcase appears. llvm-svn: 347632
2018-11-27 10:57:27 +08:00
; AVX12-NEXT: vmovmskps %xmm0, %eax
; AVX12-NEXT: movb %al, (%rdi)
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
; AVX12-NEXT: retq
;
; AVX512F-LABEL: bitcast_4i32_store:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512F-NEXT: vpcmpgtd %xmm0, %xmm1, %k0
; AVX512F-NEXT: kmovw %k0, %eax
; AVX512F-NEXT: movb %al, (%rdi)
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: bitcast_4i32_store:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512BW-NEXT: vpcmpgtd %xmm0, %xmm1, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
; AVX512BW-NEXT: movb %al, (%rdi)
; AVX512BW-NEXT: retq
%a1 = icmp slt <4 x i32> %a0, zeroinitializer
%a2 = bitcast <4 x i1> %a1 to i4
store i4 %a2, i4* %p
ret void
}
define void @bitcast_2i64_store(i2* %p, <2 x i64> %a0) {
; SSE2-SSSE3-LABEL: bitcast_2i64_store:
; SSE2-SSSE3: # %bb.0:
[X86] Prevent DAG combine from folding a bitcast from vXi1 to iX with a store on pre-AVX512 targets. If we fold the bitcast into the store we'll end up creating a truncating store to vXi1 that will get scalarized. Instead allow the bitcast to be turned into a movmsk. We probably need to do something if the store itself is a vXi1 type, but I'll leave that til a testcase appears. llvm-svn: 347632
2018-11-27 10:57:27 +08:00
; SSE2-SSSE3-NEXT: movmskpd %xmm0, %eax
; SSE2-SSSE3-NEXT: movb %al, (%rdi)
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
; SSE2-SSSE3-NEXT: retq
;
; AVX12-LABEL: bitcast_2i64_store:
; AVX12: # %bb.0:
[X86] Prevent DAG combine from folding a bitcast from vXi1 to iX with a store on pre-AVX512 targets. If we fold the bitcast into the store we'll end up creating a truncating store to vXi1 that will get scalarized. Instead allow the bitcast to be turned into a movmsk. We probably need to do something if the store itself is a vXi1 type, but I'll leave that til a testcase appears. llvm-svn: 347632
2018-11-27 10:57:27 +08:00
; AVX12-NEXT: vmovmskpd %xmm0, %eax
; AVX12-NEXT: movb %al, (%rdi)
[X86] Add a bunch of test cases for storing a scalar bitcasted from a vXi1 type. Currently a store combine will absorb the bitcast before our combine that turns bitcasts into movmsk gets a chance to run. This results in a store being created with a vXi1 type. Type legalization then promotes the input type and makes this a truncating store. Then we badly scalarize this store. Currently we avoid this on v8i1->i8 bitcasts due to an incompletely qualified(per the original intention) check in isLoadBitCastBeneficial. An easy fix is to disable this for all vXi1->iX bitcasts on pre-avx512 targets. We'll still generate terrible code if the IR explicitly contains a store of vXi1 without a bitcast. We could probably solve that by just turning all stores of vXi1 into (store (iX (bitcast))) as an early DAG combine. llvm-svn: 347631
2018-11-27 10:57:23 +08:00
; AVX12-NEXT: retq
;
; AVX512F-LABEL: bitcast_2i64_store:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512F-NEXT: vpcmpgtq %xmm0, %xmm1, %k0
; AVX512F-NEXT: kmovw %k0, %eax
; AVX512F-NEXT: movb %al, (%rdi)
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: bitcast_2i64_store:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512BW-NEXT: vpcmpgtq %xmm0, %xmm1, %k0
; AVX512BW-NEXT: kmovd %k0, %eax
; AVX512BW-NEXT: movb %al, (%rdi)
; AVX512BW-NEXT: retq
%a1 = icmp slt <2 x i64> %a0, zeroinitializer
%a2 = bitcast <2 x i1> %a1 to i2
store i2 %a2, i2* %p
ret void
}