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/PowerPC/vec_mul.ll

91 lines
3.4 KiB
LLVM
Raw Normal View History

PowerPC: Do not use llc -march in tests. `llc -march` is problematic because it only switches the target architecture, but leaves the operating system unchanged. This occasionally leads to indeterministic tests because the OS from LLVM_DEFAULT_TARGET_TRIPLE is used. However we can simply always use `llc -mtriple` instead. This changes all the tests to do this to avoid people using -march when they copy and paste parts of tests. This patch: - Removes -march if the .ll file already has a matching `target triple` directive or -mtriple argument. - In all other cases changes -march=ppc32/-march=ppc64 to -mtriple=ppc32--/-mtriple=ppc64-- See also the discussion in https://reviews.llvm.org/D35287 llvm-svn: 309754
2017-08-02 06:20:41 +08:00
; RUN: llc -verify-machineinstrs < %s -mtriple=powerpc-unknown-linux-gnu -mattr=+altivec -mattr=-vsx -mattr=-power8-altivec | FileCheck %s
; RUN: llc -verify-machineinstrs < %s -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec -mattr=-vsx -mcpu=pwr7 | FileCheck %s
; RUN: llc -verify-machineinstrs < %s -mtriple=powerpc64le-unknown-linux-gnu -mattr=+altivec -mattr=-vsx -mcpu=pwr8 -mattr=-power8-altivec | FileCheck %s -check-prefix=CHECK-LE
; RUN: llc -verify-machineinstrs < %s -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec -mattr=+vsx -mcpu=pwr7 | FileCheck %s -check-prefix=CHECK-VSX
; RUN: llc -verify-machineinstrs < %s -mtriple=powerpc64le-unknown-linux-gnu -mattr=+altivec -mattr=+vsx -mcpu=pwr8 -mattr=-power8-altivec | FileCheck %s -check-prefix=CHECK-LE-VSX
new testcase llvm-svn: 27787
2006-04-18 11:22:16 +08:00
Remove llvm-upgrade and update tests. llvm-svn: 47325
2008-02-19 16:07:33 +08:00
define <4 x i32> @test_v4i32(<4 x i32>* %X, <4 x i32>* %Y) {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp = load <4 x i32>, <4 x i32>* %X ; <<4 x i32>> [#uses=1]
%tmp2 = load <4 x i32>, <4 x i32>* %Y ; <<4 x i32>> [#uses=1]
Remove llvm-upgrade and update tests. llvm-svn: 47325
2008-02-19 16:07:33 +08:00
%tmp3 = mul <4 x i32> %tmp, %tmp2 ; <<4 x i32>> [#uses=1]
ret <4 x i32> %tmp3
new testcase llvm-svn: 27787
2006-04-18 11:22:16 +08:00
}
Convert CodeGen/*/*.ll tests to use the new CHECK-LABEL for easier debugging. No functionality change and all tests pass after conversion. This was done with the following sed invocation to catch label lines demarking function boundaries: sed -i '' "s/^;\( *\)\([A-Z0-9_]*\):\( *\)test\([A-Za-z0-9_-]*\):\( *\)$/;\1\2-LABEL:\3test\4:\5/g" test/CodeGen/*/*.ll which was written conservatively to avoid false positives rather than false negatives. I scanned through all the changes and everything looks correct. llvm-svn: 186258
2013-07-14 04:38:47 +08:00
; CHECK-LABEL: test_v4i32:
This patch fixes the Altivec addend construction for the fused multiply-add instruction (vmaddfp) to conform with IEEE to ensure the sign of a zero result when resulting product is -0.0. The -0.0 vector addend to vmaddfp is generated by a creating a vector with full bits sets and then shifting each elements by 31-bits to the left, resulting in a vector of 0x80000000 (or -0.0 as float). The 'buildvec_canonicalize.ll' was adjusted to reflect this change and the 'vec_mul.ll' was complemented with the float vector multiplication test. llvm-svn: 168998
2012-11-30 21:05:44 +08:00
; CHECK: vmsumuhm
; CHECK-NOT: mullw
[PPC64LE] Generate correct little-endian code for v16i8 multiply The existing code in PPCTargetLowering::LowerMUL() for multiplying two v16i8 values assumes that vector elements are numbered in big-endian order. For little-endian targets, the vector element numbering is reversed, but the vmuleub, vmuloub, and vperm instructions still assume big-endian numbering. To account for this, we must adjust the permute control vector and reverse the order of the input registers on the vperm instruction. The existing test/CodeGen/PowerPC/vec_mul.ll is updated to be executed on powerpc64 and powerpc64le targets as well as the original powerpc (32-bit) target. llvm-svn: 210474
2014-06-10 00:06:29 +08:00
; CHECK-LE-LABEL: test_v4i32:
; CHECK-LE: vmsumuhm
; CHECK-LE-NOT: mullw
[PPC] Adjust some PowerPC tests to account for presence/absence of VSX Patch by Bill Seurer; committed on his behalf. These test cases generate slightly different code sequences when VSX is activated and thus fail. The update turns off VSX explicitly for the existing checks and then adds a second set of checks for most of them that test the VSX instruction output. llvm-svn: 220019
2014-10-17 09:41:22 +08:00
; CHECK-VSX-LABEL: test_v4i32:
; CHECK-VSX: vmsumuhm
; CHECK-VSX-NOT: mullw
; CHECK-LE-VSX-LABEL: test_v4i32:
; CHECK-LE-VSX: vmsumuhm
; CHECK-LE-VSX-NOT: mullw
new testcase llvm-svn: 27787
2006-04-18 11:22:16 +08:00
Remove llvm-upgrade and update tests. llvm-svn: 47325
2008-02-19 16:07:33 +08:00
define <8 x i16> @test_v8i16(<8 x i16>* %X, <8 x i16>* %Y) {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp = load <8 x i16>, <8 x i16>* %X ; <<8 x i16>> [#uses=1]
%tmp2 = load <8 x i16>, <8 x i16>* %Y ; <<8 x i16>> [#uses=1]
Remove llvm-upgrade and update tests. llvm-svn: 47325
2008-02-19 16:07:33 +08:00
%tmp3 = mul <8 x i16> %tmp, %tmp2 ; <<8 x i16>> [#uses=1]
ret <8 x i16> %tmp3
Add tests for v8i16 and v16i8 llvm-svn: 27791
2006-04-18 11:54:50 +08:00
}
Convert CodeGen/*/*.ll tests to use the new CHECK-LABEL for easier debugging. No functionality change and all tests pass after conversion. This was done with the following sed invocation to catch label lines demarking function boundaries: sed -i '' "s/^;\( *\)\([A-Z0-9_]*\):\( *\)test\([A-Za-z0-9_-]*\):\( *\)$/;\1\2-LABEL:\3test\4:\5/g" test/CodeGen/*/*.ll which was written conservatively to avoid false positives rather than false negatives. I scanned through all the changes and everything looks correct. llvm-svn: 186258
2013-07-14 04:38:47 +08:00
; CHECK-LABEL: test_v8i16:
This patch fixes the Altivec addend construction for the fused multiply-add instruction (vmaddfp) to conform with IEEE to ensure the sign of a zero result when resulting product is -0.0. The -0.0 vector addend to vmaddfp is generated by a creating a vector with full bits sets and then shifting each elements by 31-bits to the left, resulting in a vector of 0x80000000 (or -0.0 as float). The 'buildvec_canonicalize.ll' was adjusted to reflect this change and the 'vec_mul.ll' was complemented with the float vector multiplication test. llvm-svn: 168998
2012-11-30 21:05:44 +08:00
; CHECK: vmladduhm
; CHECK-NOT: mullw
[PPC64LE] Generate correct little-endian code for v16i8 multiply The existing code in PPCTargetLowering::LowerMUL() for multiplying two v16i8 values assumes that vector elements are numbered in big-endian order. For little-endian targets, the vector element numbering is reversed, but the vmuleub, vmuloub, and vperm instructions still assume big-endian numbering. To account for this, we must adjust the permute control vector and reverse the order of the input registers on the vperm instruction. The existing test/CodeGen/PowerPC/vec_mul.ll is updated to be executed on powerpc64 and powerpc64le targets as well as the original powerpc (32-bit) target. llvm-svn: 210474
2014-06-10 00:06:29 +08:00
; CHECK-LE-LABEL: test_v8i16:
; CHECK-LE: vmladduhm
; CHECK-LE-NOT: mullw
[PPC] Adjust some PowerPC tests to account for presence/absence of VSX Patch by Bill Seurer; committed on his behalf. These test cases generate slightly different code sequences when VSX is activated and thus fail. The update turns off VSX explicitly for the existing checks and then adds a second set of checks for most of them that test the VSX instruction output. llvm-svn: 220019
2014-10-17 09:41:22 +08:00
; CHECK-VSX-LABEL: test_v8i16:
; CHECK-VSX: vmladduhm
; CHECK-VSX-NOT: mullw
; CHECK-LE-VSX-LABEL: test_v8i16:
; CHECK-LE-VSX: vmladduhm
; CHECK-LE-VSX-NOT: mullw
Add tests for v8i16 and v16i8 llvm-svn: 27791
2006-04-18 11:54:50 +08:00
Remove llvm-upgrade and update tests. llvm-svn: 47325
2008-02-19 16:07:33 +08:00
define <16 x i8> @test_v16i8(<16 x i8>* %X, <16 x i8>* %Y) {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp = load <16 x i8>, <16 x i8>* %X ; <<16 x i8>> [#uses=1]
%tmp2 = load <16 x i8>, <16 x i8>* %Y ; <<16 x i8>> [#uses=1]
Remove llvm-upgrade and update tests. llvm-svn: 47325
2008-02-19 16:07:33 +08:00
%tmp3 = mul <16 x i8> %tmp, %tmp2 ; <<16 x i8>> [#uses=1]
ret <16 x i8> %tmp3
Add tests for v8i16 and v16i8 llvm-svn: 27791
2006-04-18 11:54:50 +08:00
}
Convert CodeGen/*/*.ll tests to use the new CHECK-LABEL for easier debugging. No functionality change and all tests pass after conversion. This was done with the following sed invocation to catch label lines demarking function boundaries: sed -i '' "s/^;\( *\)\([A-Z0-9_]*\):\( *\)test\([A-Za-z0-9_-]*\):\( *\)$/;\1\2-LABEL:\3test\4:\5/g" test/CodeGen/*/*.ll which was written conservatively to avoid false positives rather than false negatives. I scanned through all the changes and everything looks correct. llvm-svn: 186258
2013-07-14 04:38:47 +08:00
; CHECK-LABEL: test_v16i8:
This patch fixes the Altivec addend construction for the fused multiply-add instruction (vmaddfp) to conform with IEEE to ensure the sign of a zero result when resulting product is -0.0. The -0.0 vector addend to vmaddfp is generated by a creating a vector with full bits sets and then shifting each elements by 31-bits to the left, resulting in a vector of 0x80000000 (or -0.0 as float). The 'buildvec_canonicalize.ll' was adjusted to reflect this change and the 'vec_mul.ll' was complemented with the float vector multiplication test. llvm-svn: 168998
2012-11-30 21:05:44 +08:00
; CHECK: vmuloub
; CHECK: vmuleub
; CHECK-NOT: mullw
[PPC64LE] Generate correct little-endian code for v16i8 multiply The existing code in PPCTargetLowering::LowerMUL() for multiplying two v16i8 values assumes that vector elements are numbered in big-endian order. For little-endian targets, the vector element numbering is reversed, but the vmuleub, vmuloub, and vperm instructions still assume big-endian numbering. To account for this, we must adjust the permute control vector and reverse the order of the input registers on the vperm instruction. The existing test/CodeGen/PowerPC/vec_mul.ll is updated to be executed on powerpc64 and powerpc64le targets as well as the original powerpc (32-bit) target. llvm-svn: 210474
2014-06-10 00:06:29 +08:00
; CHECK-LE-LABEL: test_v16i8:
; CHECK-LE: vmuloub [[REG1:[0-9]+]]
; CHECK-LE: vmuleub [[REG2:[0-9]+]]
; CHECK-LE: vperm {{[0-9]+}}, [[REG2]], [[REG1]]
; CHECK-LE-NOT: mullw
[PPC] Adjust some PowerPC tests to account for presence/absence of VSX Patch by Bill Seurer; committed on his behalf. These test cases generate slightly different code sequences when VSX is activated and thus fail. The update turns off VSX explicitly for the existing checks and then adds a second set of checks for most of them that test the VSX instruction output. llvm-svn: 220019
2014-10-17 09:41:22 +08:00
; CHECK-VSX-LABEL: test_v16i8:
; CHECK-VSX: vmuloub
; CHECK-VSX: vmuleub
; CHECK-VSX-NOT: mullw
; CHECK-LE-VSX-LABEL: test_v16i8:
; CHECK-LE-VSX: vmuloub [[REG1:[0-9]+]]
; CHECK-LE-VSX: vmuleub [[REG2:[0-9]+]]
; CHECK-LE-VSX: vperm {{[0-9]+}}, [[REG2]], [[REG1]]
; CHECK-LE-VSX-NOT: mullw
This patch fixes the Altivec addend construction for the fused multiply-add instruction (vmaddfp) to conform with IEEE to ensure the sign of a zero result when resulting product is -0.0. The -0.0 vector addend to vmaddfp is generated by a creating a vector with full bits sets and then shifting each elements by 31-bits to the left, resulting in a vector of 0x80000000 (or -0.0 as float). The 'buildvec_canonicalize.ll' was adjusted to reflect this change and the 'vec_mul.ll' was complemented with the float vector multiplication test. llvm-svn: 168998
2012-11-30 21:05:44 +08:00
define <4 x float> @test_float(<4 x float>* %X, <4 x float>* %Y) {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp = load <4 x float>, <4 x float>* %X
%tmp2 = load <4 x float>, <4 x float>* %Y
This patch fixes the Altivec addend construction for the fused multiply-add instruction (vmaddfp) to conform with IEEE to ensure the sign of a zero result when resulting product is -0.0. The -0.0 vector addend to vmaddfp is generated by a creating a vector with full bits sets and then shifting each elements by 31-bits to the left, resulting in a vector of 0x80000000 (or -0.0 as float). The 'buildvec_canonicalize.ll' was adjusted to reflect this change and the 'vec_mul.ll' was complemented with the float vector multiplication test. llvm-svn: 168998
2012-11-30 21:05:44 +08:00
%tmp3 = fmul <4 x float> %tmp, %tmp2
ret <4 x float> %tmp3
}
; Check the creation of a negative zero float vector by creating a vector of
; all bits set and shifting it 31 bits to left, resulting a an vector of
; 4 x 0x80000000 (-0.0 as float).
Convert CodeGen/*/*.ll tests to use the new CHECK-LABEL for easier debugging. No functionality change and all tests pass after conversion. This was done with the following sed invocation to catch label lines demarking function boundaries: sed -i '' "s/^;\( *\)\([A-Z0-9_]*\):\( *\)test\([A-Za-z0-9_-]*\):\( *\)$/;\1\2-LABEL:\3test\4:\5/g" test/CodeGen/*/*.ll which was written conservatively to avoid false positives rather than false negatives. I scanned through all the changes and everything looks correct. llvm-svn: 186258
2013-07-14 04:38:47 +08:00
; CHECK-LABEL: test_float:
test/CodeGen/PowerPC/vec_mul.ll: Add a triple. Thanks, Hal. llvm-svn: 169026
2012-12-01 03:15:10 +08:00
; CHECK: vspltisw [[ZNEG:[0-9]+]], -1
; CHECK: vslw {{[0-9]+}}, [[ZNEG]], [[ZNEG]]
This patch fixes the Altivec addend construction for the fused multiply-add instruction (vmaddfp) to conform with IEEE to ensure the sign of a zero result when resulting product is -0.0. The -0.0 vector addend to vmaddfp is generated by a creating a vector with full bits sets and then shifting each elements by 31-bits to the left, resulting in a vector of 0x80000000 (or -0.0 as float). The 'buildvec_canonicalize.ll' was adjusted to reflect this change and the 'vec_mul.ll' was complemented with the float vector multiplication test. llvm-svn: 168998
2012-11-30 21:05:44 +08:00
; CHECK: vmaddfp
[PPC64LE] Generate correct little-endian code for v16i8 multiply The existing code in PPCTargetLowering::LowerMUL() for multiplying two v16i8 values assumes that vector elements are numbered in big-endian order. For little-endian targets, the vector element numbering is reversed, but the vmuleub, vmuloub, and vperm instructions still assume big-endian numbering. To account for this, we must adjust the permute control vector and reverse the order of the input registers on the vperm instruction. The existing test/CodeGen/PowerPC/vec_mul.ll is updated to be executed on powerpc64 and powerpc64le targets as well as the original powerpc (32-bit) target. llvm-svn: 210474
2014-06-10 00:06:29 +08:00
; CHECK-LE-LABEL: test_float:
; CHECK-LE: vspltisw [[ZNEG:[0-9]+]], -1
; CHECK-LE: vslw {{[0-9]+}}, [[ZNEG]], [[ZNEG]]
; CHECK-LE: vmaddfp
[PPC] Adjust some PowerPC tests to account for presence/absence of VSX Patch by Bill Seurer; committed on his behalf. These test cases generate slightly different code sequences when VSX is activated and thus fail. The update turns off VSX explicitly for the existing checks and then adds a second set of checks for most of them that test the VSX instruction output. llvm-svn: 220019
2014-10-17 09:41:22 +08:00
; CHECK-VSX-LABEL: test_float:
; CHECK-VSX: xvmulsp
; CHECK-LE-VSX-LABEL: test_float:
; CHECK-LE-VSX: xvmulsp