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/qpx-sel.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

152 lines
3.9 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 -mcpu=a2q | FileCheck %s
[PowerPC] Add triples to QPX tests Some of these tests fail on Darwin systems because of a lack of a triple; fix that. llvm-svn: 230421
2015-02-25 09:26:59 +08:00
target triple = "powerpc64-bgq-linux"
[PowerPC] Add support for the QPX vector instruction set This adds support for the QPX vector instruction set, which is used by the enhanced A2 cores on the IBM BG/Q supercomputers. QPX vectors are 256 bytes wide, holding 4 double-precision floating-point values. Boolean values, modeled here as <4 x i1> are actually also represented as floating-point values (essentially { -1, 1 } for { false, true }). QPX shares many features with Altivec and VSX, but is distinct from both of them. One major difference is that, instead of adding completely-separate vector registers, QPX vector registers are extensions of the scalar floating-point registers (lane 0 is the corresponding scalar floating-point value). The operations supported on QPX vectors mirrors that supported on the scalar floating-point values (with some additional ones for permutations and logical/comparison operations). I've been maintaining this support out-of-tree, as part of the bgclang project, for several years. This is not the entire bgclang patch set, but is most of the subset that can be cleanly integrated into LLVM proper at this time. Adding this to the LLVM backend is part of my efforts to rebase bgclang to the current LLVM trunk, but is independently useful (especially for codes that use LLVM as a JIT in library form). The assembler/disassembler test coverage is complete. The CodeGen test coverage is not, but I've included some tests, and more will be added as follow-up work. llvm-svn: 230413
2015-02-25 09:06:45 +08:00
@R = global <4 x i1> <i1 0, i1 0, i1 0, i1 0>, align 16
define <4 x double> @test1(<4 x double> %a, <4 x double> %b, <4 x i1> %c) nounwind readnone {
entry:
%r = select <4 x i1> %c, <4 x double> %a, <4 x double> %b
ret <4 x double> %r
; CHECK-LABEL: @test1
; CHECK: qvfsel 1, 3, 1, 2
; CHECK: blr
}
define <4 x double> @test2(<4 x double> %a, <4 x double> %b, i1 %c1, i1 %c2, i1 %c3, i1 %c4) nounwind readnone {
entry:
%v = insertelement <4 x i1> undef, i1 %c1, i32 0
%v2 = insertelement <4 x i1> %v, i1 %c2, i32 1
%v3 = insertelement <4 x i1> %v2, i1 %c3, i32 2
%v4 = insertelement <4 x i1> %v3, i1 %c4, i32 3
%r = select <4 x i1> %v4, <4 x double> %a, <4 x double> %b
ret <4 x double> %r
; CHECK-LABEL: @test2
; FIXME: This load/store sequence is unnecessary.
; CHECK-DAG: lbz
; CHECK-DAG: stw
; CHECK-DAG: qvlfiwzx [[REG1:[0-9]+]],
; CHECK-DAG: qvlfdx [[REG2:[0-9]+]],
; CHECK-DAG: qvfcfidu [[REG3:[0-9]+]], [[REG1]]
; CHECK: qvfcmpeq [[REG4:[0-9]+]], [[REG3]], [[REG2]]
; CHECK: qvfsel 1, [[REG4]], 1, 2
; CHECK: blr
}
define <4 x i1> @test3(<4 x i1> %a) nounwind readnone {
entry:
%v = and <4 x i1> %a, <i1 0, i1 undef, i1 1, i1 1>
ret <4 x i1> %v
; CHECK-LABEL: @test3
; CHECK: qvlfsx [[REG:[0-9]+]],
; qvflogical 1, 1, [[REG]], 1
; blr
}
[PPC] Enable transformations in PPCPassConfig::addIRPasses at O2 http://reviews.llvm.org/D18562 A large number of testcases has been modified so they pass after this test. One testcase is deleted, because I realized even after undoing the original change that was committed with this testcase, the testcase still passes. So I removed it. The change to one other testcase (test/CodeGen/PowerPC/pr25802.ll) is an arbitrary change to keep it passing. Given the original intention of the testcase, and the fact that fixing it will require some time to change the testcase, we concluded that this quick change will be enough. llvm-svn: 265683
2016-04-07 23:30:55 +08:00
define <4 x i1> @test4(<4 x i1> %a, <4 x i1>* %t) nounwind {
[PowerPC] Add support for the QPX vector instruction set This adds support for the QPX vector instruction set, which is used by the enhanced A2 cores on the IBM BG/Q supercomputers. QPX vectors are 256 bytes wide, holding 4 double-precision floating-point values. Boolean values, modeled here as <4 x i1> are actually also represented as floating-point values (essentially { -1, 1 } for { false, true }). QPX shares many features with Altivec and VSX, but is distinct from both of them. One major difference is that, instead of adding completely-separate vector registers, QPX vector registers are extensions of the scalar floating-point registers (lane 0 is the corresponding scalar floating-point value). The operations supported on QPX vectors mirrors that supported on the scalar floating-point values (with some additional ones for permutations and logical/comparison operations). I've been maintaining this support out-of-tree, as part of the bgclang project, for several years. This is not the entire bgclang patch set, but is most of the subset that can be cleanly integrated into LLVM proper at this time. Adding this to the LLVM backend is part of my efforts to rebase bgclang to the current LLVM trunk, but is independently useful (especially for codes that use LLVM as a JIT in library form). The assembler/disassembler test coverage is complete. The CodeGen test coverage is not, but I've included some tests, and more will be added as follow-up work. llvm-svn: 230413
2015-02-25 09:06:45 +08:00
entry:
[PPC] Enable transformations in PPCPassConfig::addIRPasses at O2 http://reviews.llvm.org/D18562 A large number of testcases has been modified so they pass after this test. One testcase is deleted, because I realized even after undoing the original change that was committed with this testcase, the testcase still passes. So I removed it. The change to one other testcase (test/CodeGen/PowerPC/pr25802.ll) is an arbitrary change to keep it passing. Given the original intention of the testcase, and the fact that fixing it will require some time to change the testcase, we concluded that this quick change will be enough. llvm-svn: 265683
2016-04-07 23:30:55 +08:00
%q = load <4 x i1>, <4 x i1>* %t, align 16
[PowerPC] Add support for the QPX vector instruction set This adds support for the QPX vector instruction set, which is used by the enhanced A2 cores on the IBM BG/Q supercomputers. QPX vectors are 256 bytes wide, holding 4 double-precision floating-point values. Boolean values, modeled here as <4 x i1> are actually also represented as floating-point values (essentially { -1, 1 } for { false, true }). QPX shares many features with Altivec and VSX, but is distinct from both of them. One major difference is that, instead of adding completely-separate vector registers, QPX vector registers are extensions of the scalar floating-point registers (lane 0 is the corresponding scalar floating-point value). The operations supported on QPX vectors mirrors that supported on the scalar floating-point values (with some additional ones for permutations and logical/comparison operations). I've been maintaining this support out-of-tree, as part of the bgclang project, for several years. This is not the entire bgclang patch set, but is most of the subset that can be cleanly integrated into LLVM proper at this time. Adding this to the LLVM backend is part of my efforts to rebase bgclang to the current LLVM trunk, but is independently useful (especially for codes that use LLVM as a JIT in library form). The assembler/disassembler test coverage is complete. The CodeGen test coverage is not, but I've included some tests, and more will be added as follow-up work. llvm-svn: 230413
2015-02-25 09:06:45 +08:00
%v = and <4 x i1> %a, %q
ret <4 x i1> %v
; CHECK-LABEL: @test4
; CHECK-DAG: lbz
; CHECK-DAG: qvlfdx [[REG1:[0-9]+]],
; CHECK-DAG: stw
; CHECK-DAG: qvlfiwzx [[REG2:[0-9]+]],
; CHECK-DAG: qvfcfidu [[REG3:[0-9]+]], [[REG2]]
; CHECK: qvfcmpeq [[REG4:[0-9]+]], [[REG3]], [[REG1]]
; CHECK: qvflogical 1, 1, [[REG4]], 1
; CHECK: blr
}
define void @test5(<4 x i1> %a) nounwind {
entry:
store <4 x i1> %a, <4 x i1>* @R
ret void
; CHECK-LABEL: @test5
; CHECK: qvlfdx [[REG1:[0-9]+]],
; CHECK: qvfmadd [[REG2:[0-9]+]], 1, [[REG1]], [[REG1]]
; CHECK: qvfctiwu [[REG3:[0-9]+]], [[REG2]]
; CHECK: qvstfiwx [[REG3]],
; CHECK: lwz
; CHECK: stb
; CHECK: blr
}
define i1 @test6(<4 x i1> %a) nounwind {
entry:
%r = extractelement <4 x i1> %a, i32 2
ret i1 %r
; CHECK-LABEL: @test6
; CHECK: qvlfdx [[REG1:[0-9]+]],
; CHECK: qvfmadd [[REG2:[0-9]+]], 1, [[REG1]], [[REG1]]
; CHECK: qvfctiwu [[REG3:[0-9]+]], [[REG2]]
; CHECK: qvstfiwx [[REG3]],
; CHECK: lwz
; CHECK: blr
}
define i1 @test7(<4 x i1> %a) nounwind {
entry:
%r = extractelement <4 x i1> %a, i32 2
%s = extractelement <4 x i1> %a, i32 3
%q = and i1 %r, %s
ret i1 %q
; CHECK-LABEL: @test7
; CHECK: qvlfdx [[REG1:[0-9]+]],
; CHECK: qvfmadd [[REG2:[0-9]+]], 1, [[REG1]], [[REG1]]
; CHECK: qvfctiwu [[REG3:[0-9]+]], [[REG2]]
; CHECK: qvstfiwx [[REG3]],
; CHECK-DAG: lwz [[REG4:[0-9]+]],
; FIXME: We're storing the vector twice, and that's silly.
; CHECK-DAG: qvstfiwx [[REG3]],
; CHECK-DAG: lwz [[REG5:[0-9]+]],
; CHECK: and 3,
; CHECK: blr
}
define i1 @test8(<3 x i1> %a) nounwind {
entry:
%r = extractelement <3 x i1> %a, i32 2
ret i1 %r
; CHECK-LABEL: @test8
; CHECK: qvlfdx [[REG1:[0-9]+]],
; CHECK: qvfmadd [[REG2:[0-9]+]], 1, [[REG1]], [[REG1]]
; CHECK: qvfctiwu [[REG3:[0-9]+]], [[REG2]]
; CHECK: qvstfiwx [[REG3]],
; CHECK: lwz
; CHECK: blr
}
define <3 x double> @test9(<3 x double> %a, <3 x double> %b, i1 %c1, i1 %c2, i1 %c3) nounwind readnone {
entry:
%v = insertelement <3 x i1> undef, i1 %c1, i32 0
%v2 = insertelement <3 x i1> %v, i1 %c2, i32 1
%v3 = insertelement <3 x i1> %v2, i1 %c3, i32 2
%r = select <3 x i1> %v3, <3 x double> %a, <3 x double> %b
ret <3 x double> %r
; CHECK-LABEL: @test9
; FIXME: This load/store sequence is unnecessary.
; CHECK-DAG: lbz
; CHECK-DAG: stw
; CHECK-DAG: qvlfiwzx [[REG1:[0-9]+]],
; CHECK-DAG: qvlfdx [[REG2:[0-9]+]],
; CHECK-DAG: qvfcfidu [[REG3:[0-9]+]], [[REG1]]
; CHECK: qvfcmpeq [[REG4:[0-9]+]], [[REG3]], [[REG2]]
; CHECK: qvfsel 1, [[REG4]], 1, 2
; CHECK: blr
}