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/avx512f-vec-test-testn.ll

180 lines
6.3 KiB
LLVM
Raw Normal View History

[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s
; Function Attrs: norecurse nounwind readnone
define zeroext i8 @TEST_mm512_test_epi64_mask(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_test_epi64_mask:
[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
; CHECK: # %bb.0: # %entry
[X86][AVX512] Improve lowering of AVX512 test intrinsics Added TESTM and TESTNM to the list of instructions that already zeroing unused upper bits and does not need the redundant shift left and shift right instructions afterwards. Added a pattern for TESTM and TESTNM in iselLowering, so now icmp(neq,and(X,Y), 0) goes folds into TESTM and icmp(eq,and(X,Y), 0) goes folds into TESTNM This commit is a preparation for lowering the test and testn X86 intrinsics to IR. Differential Revision: https://reviews.llvm.org/D38732 llvm-svn: 317465
2017-11-06 17:22:38 +08:00
; CHECK-NEXT: vptestmq %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-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
; CHECK-NEXT: # kill: def $al killed $al killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = icmp ne <8 x i64> %and.i.i, zeroinitializer
%1 = bitcast <8 x i1> %0 to i8
ret i8 %1
}
[X86] Reduce the number of isel pattern variations needed for VPTESTM/VPTESTNM matching. Canonicalize EQ/NE PCMPM to have build vector all zeros on the RHS so we don't have to pattern match it in both locations. This significantly reduces the number of isel patterns needed since we also had to multiply it out with loads being in either operand of the 'and' input node and in the 'and' masking node. This removes over 24000 bytes from the isel table. llvm-svn: 325526
2018-02-20 03:23:31 +08:00
; Similar to the above, but the compare is reversed to have the zeros on the LHS
define zeroext i8 @TEST_mm512_test_epi64_mask_2(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_test_epi64_mask_2:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vptestmq %zmm0, %zmm1, %k0
; CHECK-NEXT: kmovw %k0, %eax
; CHECK-NEXT: # kill: def $al killed $al killed $eax
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = icmp ne <8 x i64> zeroinitializer, %and.i.i
%1 = bitcast <8 x i1> %0 to i8
ret i8 %1
}
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; Function Attrs: norecurse nounwind readnone
define zeroext i16 @TEST_mm512_test_epi32_mask(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_test_epi32_mask:
[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
; CHECK: # %bb.0: # %entry
[X86][AVX512] Improve lowering of AVX512 test intrinsics Added TESTM and TESTNM to the list of instructions that already zeroing unused upper bits and does not need the redundant shift left and shift right instructions afterwards. Added a pattern for TESTM and TESTNM in iselLowering, so now icmp(neq,and(X,Y), 0) goes folds into TESTM and icmp(eq,and(X,Y), 0) goes folds into TESTNM This commit is a preparation for lowering the test and testn X86 intrinsics to IR. Differential Revision: https://reviews.llvm.org/D38732 llvm-svn: 317465
2017-11-06 17:22:38 +08:00
; CHECK-NEXT: vptestmd %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-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
; CHECK-NEXT: # kill: def $ax killed $ax killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = bitcast <8 x i64> %and.i.i to <16 x i32>
%1 = icmp ne <16 x i32> %0, zeroinitializer
%2 = bitcast <16 x i1> %1 to i16
ret i16 %2
}
; Function Attrs: norecurse nounwind readnone
define zeroext i8 @TEST_mm512_mask_test_epi64_mask(i8 %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_mask_test_epi64_mask:
[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
; CHECK: # %bb.0: # %entry
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: vptestmq %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: kmovw %k0, %eax
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: andb %dil, %al
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
; CHECK-NEXT: # kill: def $al killed $al killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = icmp ne <8 x i64> %and.i.i, zeroinitializer
%1 = bitcast i8 %__U to <8 x i1>
%2 = and <8 x i1> %0, %1
%3 = bitcast <8 x i1> %2 to i8
ret i8 %3
}
; Function Attrs: norecurse nounwind readnone
define zeroext i16 @TEST_mm512_mask_test_epi32_mask(i16 %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_mask_test_epi32_mask:
[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
; CHECK: # %bb.0: # %entry
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: vptestmd %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: kmovw %k0, %eax
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: andl %edi, %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
; CHECK-NEXT: # kill: def $ax killed $ax killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = bitcast <8 x i64> %and.i.i to <16 x i32>
%1 = icmp ne <16 x i32> %0, zeroinitializer
%2 = bitcast i16 %__U to <16 x i1>
%3 = and <16 x i1> %1, %2
%4 = bitcast <16 x i1> %3 to i16
ret i16 %4
}
; Function Attrs: norecurse nounwind readnone
define zeroext i8 @TEST_mm512_testn_epi64_mask(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_testn_epi64_mask:
[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
; CHECK: # %bb.0: # %entry
[X86][AVX512] Improve lowering of AVX512 test intrinsics Added TESTM and TESTNM to the list of instructions that already zeroing unused upper bits and does not need the redundant shift left and shift right instructions afterwards. Added a pattern for TESTM and TESTNM in iselLowering, so now icmp(neq,and(X,Y), 0) goes folds into TESTM and icmp(eq,and(X,Y), 0) goes folds into TESTNM This commit is a preparation for lowering the test and testn X86 intrinsics to IR. Differential Revision: https://reviews.llvm.org/D38732 llvm-svn: 317465
2017-11-06 17:22:38 +08:00
; CHECK-NEXT: vptestnmq %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-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
; CHECK-NEXT: # kill: def $al killed $al killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = icmp eq <8 x i64> %and.i.i, zeroinitializer
%1 = bitcast <8 x i1> %0 to i8
ret i8 %1
}
[X86] Reduce the number of isel pattern variations needed for VPTESTM/VPTESTNM matching. Canonicalize EQ/NE PCMPM to have build vector all zeros on the RHS so we don't have to pattern match it in both locations. This significantly reduces the number of isel patterns needed since we also had to multiply it out with loads being in either operand of the 'and' input node and in the 'and' masking node. This removes over 24000 bytes from the isel table. llvm-svn: 325526
2018-02-20 03:23:31 +08:00
; Similar to the above, but the compare is reversed to have the zeros on the LHS
define zeroext i8 @TEST_mm512_testn_epi64_mask_2(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_testn_epi64_mask_2:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vptestnmq %zmm0, %zmm1, %k0
; CHECK-NEXT: kmovw %k0, %eax
; CHECK-NEXT: # kill: def $al killed $al killed $eax
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = icmp eq <8 x i64> zeroinitializer, %and.i.i
%1 = bitcast <8 x i1> %0 to i8
ret i8 %1
}
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; Function Attrs: norecurse nounwind readnone
define zeroext i16 @TEST_mm512_testn_epi32_mask(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_testn_epi32_mask:
[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
; CHECK: # %bb.0: # %entry
[X86][AVX512] Improve lowering of AVX512 test intrinsics Added TESTM and TESTNM to the list of instructions that already zeroing unused upper bits and does not need the redundant shift left and shift right instructions afterwards. Added a pattern for TESTM and TESTNM in iselLowering, so now icmp(neq,and(X,Y), 0) goes folds into TESTM and icmp(eq,and(X,Y), 0) goes folds into TESTNM This commit is a preparation for lowering the test and testn X86 intrinsics to IR. Differential Revision: https://reviews.llvm.org/D38732 llvm-svn: 317465
2017-11-06 17:22:38 +08:00
; CHECK-NEXT: vptestnmd %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-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
; CHECK-NEXT: # kill: def $ax killed $ax killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = bitcast <8 x i64> %and.i.i to <16 x i32>
%1 = icmp eq <16 x i32> %0, zeroinitializer
%2 = bitcast <16 x i1> %1 to i16
ret i16 %2
}
; Function Attrs: norecurse nounwind readnone
define zeroext i8 @TEST_mm512_mask_testn_epi64_mask(i8 %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_mask_testn_epi64_mask:
[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
; CHECK: # %bb.0: # %entry
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: vptestnmq %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: kmovw %k0, %eax
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: andb %dil, %al
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
; CHECK-NEXT: # kill: def $al killed $al killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = icmp eq <8 x i64> %and.i.i, zeroinitializer
%1 = bitcast i8 %__U to <8 x i1>
%2 = and <8 x i1> %0, %1
%3 = bitcast <8 x i1> %2 to i8
ret i8 %3
}
; Function Attrs: norecurse nounwind readnone
define zeroext i16 @TEST_mm512_mask_testn_epi32_mask(i16 %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
; CHECK-LABEL: TEST_mm512_mask_testn_epi32_mask:
[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
; CHECK: # %bb.0: # %entry
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: vptestnmd %zmm0, %zmm1, %k0
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: kmovw %k0, %eax
[X86] Add DAG combine to turn (bitcast (and/or/xor (bitcast X), Y)) -> (and/or/xor X, (bitcast Y)) when casting between GPRs and mask operations. This reduces the number of transitions between k-registers and GPRs, reducing the number of instructions. There's still some room for improvement to remove more transitions, but this is a good start. llvm-svn: 324184
2018-02-04 09:43:48 +08:00
; CHECK-NEXT: andl %edi, %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
; CHECK-NEXT: # kill: def $ax killed $ax killed $eax
[X86] Added tests for TESTM and TESTNM (NFC) Adding this test files now so after another commit that will add a new pattern for TESTM and TESTNM instructions will show the improvemnts that have been done. Change-Id: If3908b7f91897d764053312365a2bc1de78b291d llvm-svn: 315443
2017-10-11 16:39:25 +08:00
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
entry:
%and.i.i = and <8 x i64> %__B, %__A
%0 = bitcast <8 x i64> %and.i.i to <16 x i32>
%1 = icmp eq <16 x i32> %0, zeroinitializer
%2 = bitcast i16 %__U to <16 x i1>
%3 = and <16 x i1> %1, %2
%4 = bitcast <16 x i1> %3 to i16
ret i16 %4
}