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/shrink-compare.ll

240 lines
4.8 KiB
LLVM
Raw Normal View History

X86: Fold tail calls into conditional branches also for 64-bit (PR26302) This extends the optimization in r280832 to also work for 64-bit. The only quirk is that we can't do this for 64-bit Windows (yet). Differential Revision: https://reviews.llvm.org/D24423 llvm-svn: 281113
2016-09-10 06:37:27 +08:00
; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s
DAGCombine: fold "(zext x) == C" into "x == (trunc C)" if the trunc is lossless. On x86 this allows to fold a load into the cmp, greatly reducing register pressure. movzbl (%rdi), %eax cmpl $47, %eax -> cmpb $47, (%rdi) This shaves 8k off gcc.o on i386. I'll leave applying the patch in README.txt to Chris :) llvm-svn: 130005
2011-04-23 02:47:44 +08:00
declare void @bar()
X86: Perform integer comparisons at i32 or larger. Utilizing the 8 and 16 bit comparison instructions, even when an input can be folded into the comparison instruction itself, is typically not worth it. There are too many partial register stalls as a result, leading to significant slowdowns. By always performing comparisons on at least 32-bit registers, performance of the calculation chain leading to the comparison improves. Continue to use the smaller comparisons when minimizing size, as that allows better folding of loads into the comparison instructions. rdar://15386341 llvm-svn: 195496
2013-11-23 03:57:47 +08:00
define void @test1(i32* nocapture %X) nounwind minsize {
DAGCombine: fold "(zext x) == C" into "x == (trunc C)" if the trunc is lossless. On x86 this allows to fold a load into the cmp, greatly reducing register pressure. movzbl (%rdi), %eax cmpl $47, %eax -> cmpb $47, (%rdi) This shaves 8k off gcc.o on i386. I'll leave applying the patch in README.txt to Chris :) llvm-svn: 130005
2011-04-23 02:47:44 +08:00
entry:
[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
%tmp1 = load i32, i32* %X, align 4
DAGCombine: fold "(zext x) == C" into "x == (trunc C)" if the trunc is lossless. On x86 this allows to fold a load into the cmp, greatly reducing register pressure. movzbl (%rdi), %eax cmpl $47, %eax -> cmpb $47, (%rdi) This shaves 8k off gcc.o on i386. I'll leave applying the patch in README.txt to Chris :) llvm-svn: 130005
2011-04-23 02:47:44 +08:00
%and = and i32 %tmp1, 255
%cmp = icmp eq i32 %and, 47
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
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: test1:
test/CodeGen/X86/shrink-compare.ll: Relax expressions for Win64. llvm-svn: 130039
2011-04-23 08:15:45 +08:00
; CHECK: cmpb $47, (%{{rdi|rcx}})
DAGCombine: fold "(zext x) == C" into "x == (trunc C)" if the trunc is lossless. On x86 this allows to fold a load into the cmp, greatly reducing register pressure. movzbl (%rdi), %eax cmpl $47, %eax -> cmpb $47, (%rdi) This shaves 8k off gcc.o on i386. I'll leave applying the patch in README.txt to Chris :) llvm-svn: 130005
2011-04-23 02:47:44 +08:00
}
X86: Perform integer comparisons at i32 or larger. Utilizing the 8 and 16 bit comparison instructions, even when an input can be folded into the comparison instruction itself, is typically not worth it. There are too many partial register stalls as a result, leading to significant slowdowns. By always performing comparisons on at least 32-bit registers, performance of the calculation chain leading to the comparison improves. Continue to use the smaller comparisons when minimizing size, as that allows better folding of loads into the comparison instructions. rdar://15386341 llvm-svn: 195496
2013-11-23 03:57:47 +08:00
define void @test2(i32 %X) nounwind minsize {
DAGCombine: fold "(zext x) == C" into "x == (trunc C)" if the trunc is lossless. On x86 this allows to fold a load into the cmp, greatly reducing register pressure. movzbl (%rdi), %eax cmpl $47, %eax -> cmpb $47, (%rdi) This shaves 8k off gcc.o on i386. I'll leave applying the patch in README.txt to Chris :) llvm-svn: 130005
2011-04-23 02:47:44 +08:00
entry:
%and = and i32 %X, 255
%cmp = icmp eq i32 %and, 47
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
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: test2:
test/CodeGen/X86/shrink-compare.ll: Relax expressions for Win64. llvm-svn: 130039
2011-04-23 08:15:45 +08:00
; CHECK: cmpb $47, %{{dil|cl}}
DAGCombine: fold "(zext x) == C" into "x == (trunc C)" if the trunc is lossless. On x86 this allows to fold a load into the cmp, greatly reducing register pressure. movzbl (%rdi), %eax cmpl $47, %eax -> cmpb $47, (%rdi) This shaves 8k off gcc.o on i386. I'll leave applying the patch in README.txt to Chris :) llvm-svn: 130005
2011-04-23 02:47:44 +08:00
}
DAGCombine: Also shrink eq compares where the constant is exactly as large as the smaller type. if ((x & 255) == 255) before: movzbl %al, %eax cmpl $255, %eax after: cmpb $-1, %al llvm-svn: 182038
2013-05-17 02:47:58 +08:00
X86: Perform integer comparisons at i32 or larger. Utilizing the 8 and 16 bit comparison instructions, even when an input can be folded into the comparison instruction itself, is typically not worth it. There are too many partial register stalls as a result, leading to significant slowdowns. By always performing comparisons on at least 32-bit registers, performance of the calculation chain leading to the comparison improves. Continue to use the smaller comparisons when minimizing size, as that allows better folding of loads into the comparison instructions. rdar://15386341 llvm-svn: 195496
2013-11-23 03:57:47 +08:00
define void @test3(i32 %X) nounwind minsize {
DAGCombine: Also shrink eq compares where the constant is exactly as large as the smaller type. if ((x & 255) == 255) before: movzbl %al, %eax cmpl $255, %eax after: cmpb $-1, %al llvm-svn: 182038
2013-05-17 02:47:58 +08:00
entry:
%and = and i32 %X, 255
%cmp = icmp eq i32 %and, 255
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
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: test3:
DAGCombine: Also shrink eq compares where the constant is exactly as large as the smaller type. if ((x & 255) == 255) before: movzbl %al, %eax cmpl $255, %eax after: cmpb $-1, %al llvm-svn: 182038
2013-05-17 02:47:58 +08:00
; CHECK: cmpb $-1, %{{dil|cl}}
}
DAGCombine: Avoid an edge case where it tried to create an i0 type for (x & 0) == 0. Fixes PR16083. llvm-svn: 182357
2013-05-21 16:51:09 +08:00
; PR16083
define i1 @test4(i64 %a, i32 %b) {
entry:
%tobool = icmp ne i32 %b, 0
br i1 %tobool, label %lor.end, label %lor.rhs
lor.rhs: ; preds = %entry
%and = and i64 0, %a
%tobool1 = icmp ne i64 %and, 0
br label %lor.end
lor.end: ; preds = %lor.rhs, %entry
%p = phi i1 [ true, %entry ], [ %tobool1, %lor.rhs ]
ret i1 %p
}
DAGCombiner: Don't drop extension behavior when shrinking a load when unsafe. ReduceLoadWidth unconditionally drops extensions from loads. Limit it to the case when all of the bits the extension would otherwise produce are dropped by the shrink. It would be possible to shrink the load in more cases by merging the extensions, but this isn't trivial and a very rare case. I left a TODO for that case. Fixes PR16551. llvm-svn: 185755
2013-07-06 22:05:09 +08:00
@x = global { i8, i8, i8, i8, i8, i8, i8, i8 } { i8 1, i8 0, i8 0, i8 0, i8 1, i8 0, i8 0, i8 1 }, align 4
; PR16551
X86: Perform integer comparisons at i32 or larger. Utilizing the 8 and 16 bit comparison instructions, even when an input can be folded into the comparison instruction itself, is typically not worth it. There are too many partial register stalls as a result, leading to significant slowdowns. By always performing comparisons on at least 32-bit registers, performance of the calculation chain leading to the comparison improves. Continue to use the smaller comparisons when minimizing size, as that allows better folding of loads into the comparison instructions. rdar://15386341 llvm-svn: 195496
2013-11-23 03:57:47 +08:00
define void @test5(i32 %X) nounwind minsize {
DAGCombiner: Don't drop extension behavior when shrinking a load when unsafe. ReduceLoadWidth unconditionally drops extensions from loads. Limit it to the case when all of the bits the extension would otherwise produce are dropped by the shrink. It would be possible to shrink the load in more cases by merging the extensions, but this isn't trivial and a very rare case. I left a TODO for that case. Fixes PR16551. llvm-svn: 185755
2013-07-06 22:05:09 +08:00
entry:
[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
%bf.load = load i56, i56* bitcast ({ i8, i8, i8, i8, i8, i8, i8, i8 }* @x to i56*), align 4
DAGCombiner: Don't drop extension behavior when shrinking a load when unsafe. ReduceLoadWidth unconditionally drops extensions from loads. Limit it to the case when all of the bits the extension would otherwise produce are dropped by the shrink. It would be possible to shrink the load in more cases by merging the extensions, but this isn't trivial and a very rare case. I left a TODO for that case. Fixes PR16551. llvm-svn: 185755
2013-07-06 22:05:09 +08:00
%bf.lshr = lshr i56 %bf.load, 32
%bf.cast = trunc i56 %bf.lshr to i32
%cmp = icmp ne i32 %bf.cast, 1
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
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: test5:
DAGCombiner: Don't drop extension behavior when shrinking a load when unsafe. ReduceLoadWidth unconditionally drops extensions from loads. Limit it to the case when all of the bits the extension would otherwise produce are dropped by the shrink. It would be possible to shrink the load in more cases by merging the extensions, but this isn't trivial and a very rare case. I left a TODO for that case. Fixes PR16551. llvm-svn: 185755
2013-07-06 22:05:09 +08:00
; CHECK-NOT: cmpl $1,{{.*}}x+4
; CHECK: ret
}
Enable (sext x) == C --> x == (trunc C) combine Extend the existing code which handles this for zext. This makes this more useful for targets with ZeroOrNegativeOne BooleanContent and obsoletes a custom combine SI uses for i1 setcc (sext(i1), 0, setne) since the constant will now be shrunk to i1. llvm-svn: 224691
2014-12-22 00:48:42 +08:00
; CHECK-LABEL: test2_1:
; CHECK: movzbl
; CHECK: cmpl $256
[X86] Re-enable conditional tail calls and fix PR31257. This reverts r294348, which removed support for conditional tail calls due to the PR above. It fixes the PR by marking live registers as implicitly used and defined by the now predicated tailcall. This is similar to how IfConversion predicates instructions. Differential Revision: https://reviews.llvm.org/D29856 llvm-svn: 295262
2017-02-16 08:04:05 +08:00
; CHECK: je bar
Enable (sext x) == C --> x == (trunc C) combine Extend the existing code which handles this for zext. This makes this more useful for targets with ZeroOrNegativeOne BooleanContent and obsoletes a custom combine SI uses for i1 setcc (sext(i1), 0, setne) since the constant will now be shrunk to i1. llvm-svn: 224691
2014-12-22 00:48:42 +08:00
define void @test2_1(i32 %X) nounwind minsize {
entry:
%and = and i32 %X, 255
%cmp = icmp eq i32 %and, 256
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_1:
; CHECK: cmpb $1, %{{dil|cl}}
define void @test_sext_i8_icmp_1(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, 1
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_47:
; CHECK: cmpb $47, %{{dil|cl}}
define void @test_sext_i8_icmp_47(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, 47
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_127:
; CHECK: cmpb $127, %{{dil|cl}}
define void @test_sext_i8_icmp_127(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, 127
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_neg1:
; CHECK: cmpb $-1, %{{dil|cl}}
define void @test_sext_i8_icmp_neg1(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, -1
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_neg2:
; CHECK: cmpb $-2, %{{dil|cl}}
define void @test_sext_i8_icmp_neg2(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, -2
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_neg127:
; CHECK: cmpb $-127, %{{dil|cl}}
define void @test_sext_i8_icmp_neg127(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, -127
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_neg128:
; CHECK: cmpb $-128, %{{dil|cl}}
define void @test_sext_i8_icmp_neg128(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, -128
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}
; CHECK-LABEL: test_sext_i8_icmp_255:
; CHECK: movb $1,
; CHECK: testb
[X86] Re-enable conditional tail calls and fix PR31257. This reverts r294348, which removed support for conditional tail calls due to the PR above. It fixes the PR by marking live registers as implicitly used and defined by the now predicated tailcall. This is similar to how IfConversion predicates instructions. Differential Revision: https://reviews.llvm.org/D29856 llvm-svn: 295262
2017-02-16 08:04:05 +08:00
; CHECK: je bar
Enable (sext x) == C --> x == (trunc C) combine Extend the existing code which handles this for zext. This makes this more useful for targets with ZeroOrNegativeOne BooleanContent and obsoletes a custom combine SI uses for i1 setcc (sext(i1), 0, setne) since the constant will now be shrunk to i1. llvm-svn: 224691
2014-12-22 00:48:42 +08:00
define void @test_sext_i8_icmp_255(i8 %x) nounwind minsize {
entry:
%sext = sext i8 %x to i32
%cmp = icmp eq i32 %sext, 255
br i1 %cmp, label %if.then, label %if.end
if.then:
tail call void @bar() nounwind
br label %if.end
if.end:
ret void
}