maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/Transforms/LoopVectorize/PowerPC/vsx-tsvc-s173.ll

52 lines
2.3 KiB
LLVM
Raw Normal View History

[LoopVectorizer] Count dependencies of consecutive pointers as uniforms For the purpose of calculating the cost of the loop at various vectorization factors, we need to count dependencies of consecutive pointers as uniforms (which means that the VF = 1 cost is used for all overall VF values). For example, the TSVC benchmark function s173 has: ... %3 = add nsw i64 %indvars.iv, 16000 %arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3 ... and we must realize that the add will be a scalar in order to correctly deduce it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all dependencies of a consecutive pointer must be a scalar (uniform), and so we simply need to add all consecutive pointers to the worklist that currently detects collects uniforms. Fixes PR19296. llvm-svn: 205387
2014-04-02 10:34:49 +08:00
; RUN: opt < %s -mcpu=pwr7 -mattr=+vsx -loop-vectorize -instcombine -S | FileCheck %s
target datalayout = "E-m:e-i64:64-n32:64"
target triple = "powerpc64-unknown-linux-gnu"
%struct.GlobalData = type { [32000 x float], [3 x i32], [4 x i8], [32000 x float], [5 x i32], [12 x i8], [32000 x float], [7 x i32], [4 x i8], [32000 x float], [11 x i32], [4 x i8], [32000 x float], [13 x i32], [12 x i8], [256 x [256 x float]], [17 x i32], [12 x i8], [256 x [256 x float]], [19 x i32], [4 x i8], [256 x [256 x float]], [23 x i32], [4 x i8], [256 x [256 x float]] }
@global_data = external global %struct.GlobalData, align 16
@ntimes = external hidden unnamed_addr global i32, align 4
define signext i32 @s173() #0 {
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
%0 = load i32, i32* @ntimes, align 4
[LoopVectorizer] Count dependencies of consecutive pointers as uniforms For the purpose of calculating the cost of the loop at various vectorization factors, we need to count dependencies of consecutive pointers as uniforms (which means that the VF = 1 cost is used for all overall VF values). For example, the TSVC benchmark function s173 has: ... %3 = add nsw i64 %indvars.iv, 16000 %arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3 ... and we must realize that the add will be a scalar in order to correctly deduce it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all dependencies of a consecutive pointer must be a scalar (uniform), and so we simply need to add all consecutive pointers to the worklist that currently detects collects uniforms. Fixes PR19296. llvm-svn: 205387
2014-04-02 10:34:49 +08:00
%cmp21 = icmp sgt i32 %0, 0
br i1 %cmp21, label %for.cond1.preheader, label %for.end12
for.cond1.preheader: ; preds = %for.end, %entry
%nl.022 = phi i32 [ %inc11, %for.end ], [ 0, %entry ]
br label %for.body3
for.body3: ; preds = %for.body3, %for.cond1.preheader
%indvars.iv = phi i64 [ 0, %for.cond1.preheader ], [ %indvars.iv.next, %for.body3 ]
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx = getelementptr inbounds %struct.GlobalData, %struct.GlobalData* @global_data, i64 0, i32 0, i64 %indvars.iv
[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
%1 = load float, float* %arrayidx, align 4
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx5 = getelementptr inbounds %struct.GlobalData, %struct.GlobalData* @global_data, i64 0, i32 3, i64 %indvars.iv
[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
%2 = load float, float* %arrayidx5, align 4
[LoopVectorizer] Count dependencies of consecutive pointers as uniforms For the purpose of calculating the cost of the loop at various vectorization factors, we need to count dependencies of consecutive pointers as uniforms (which means that the VF = 1 cost is used for all overall VF values). For example, the TSVC benchmark function s173 has: ... %3 = add nsw i64 %indvars.iv, 16000 %arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3 ... and we must realize that the add will be a scalar in order to correctly deduce it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all dependencies of a consecutive pointer must be a scalar (uniform), and so we simply need to add all consecutive pointers to the worklist that currently detects collects uniforms. Fixes PR19296. llvm-svn: 205387
2014-04-02 10:34:49 +08:00
%add = fadd float %1, %2
%3 = add nsw i64 %indvars.iv, 16000
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx8 = getelementptr inbounds %struct.GlobalData, %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3
[LoopVectorizer] Count dependencies of consecutive pointers as uniforms For the purpose of calculating the cost of the loop at various vectorization factors, we need to count dependencies of consecutive pointers as uniforms (which means that the VF = 1 cost is used for all overall VF values). For example, the TSVC benchmark function s173 has: ... %3 = add nsw i64 %indvars.iv, 16000 %arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3 ... and we must realize that the add will be a scalar in order to correctly deduce it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all dependencies of a consecutive pointer must be a scalar (uniform), and so we simply need to add all consecutive pointers to the worklist that currently detects collects uniforms. Fixes PR19296. llvm-svn: 205387
2014-04-02 10:34:49 +08:00
store float %add, float* %arrayidx8, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 16000
br i1 %exitcond, label %for.end, label %for.body3
for.end: ; preds = %for.body3
%inc11 = add nsw i32 %nl.022, 1
[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
%4 = load i32, i32* @ntimes, align 4
[LoopVectorizer] Count dependencies of consecutive pointers as uniforms For the purpose of calculating the cost of the loop at various vectorization factors, we need to count dependencies of consecutive pointers as uniforms (which means that the VF = 1 cost is used for all overall VF values). For example, the TSVC benchmark function s173 has: ... %3 = add nsw i64 %indvars.iv, 16000 %arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3 ... and we must realize that the add will be a scalar in order to correctly deduce it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all dependencies of a consecutive pointer must be a scalar (uniform), and so we simply need to add all consecutive pointers to the worklist that currently detects collects uniforms. Fixes PR19296. llvm-svn: 205387
2014-04-02 10:34:49 +08:00
%mul = mul nsw i32 %4, 10
%cmp = icmp slt i32 %inc11, %mul
br i1 %cmp, label %for.cond1.preheader, label %for.end12
for.end12: ; preds = %for.end, %entry
ret i32 0
; CHECK-LABEL: @s173
[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
; CHECK: load <4 x float>, <4 x float>*
[InstCombine] scalarizePHI should not assume the code it sees has been CSE'd scalarizePHI only looked for phis that have exactly two uses - the "latch" use, and an extract. Unfortunately, we can not assume all equivalent extracts are CSE'd, since InstCombine itself may create an extract which is a duplicate of an existing one. This extends it to handle several distinct extracts from the same index. This should fix at least some of the performance regressions from PR27988. Differential Revision: http://reviews.llvm.org/D20983 llvm-svn: 271961
2016-06-07 07:38:33 +08:00
; CHECK: add nsw i64 %1, 16000
[LoopVectorizer] Count dependencies of consecutive pointers as uniforms For the purpose of calculating the cost of the loop at various vectorization factors, we need to count dependencies of consecutive pointers as uniforms (which means that the VF = 1 cost is used for all overall VF values). For example, the TSVC benchmark function s173 has: ... %3 = add nsw i64 %indvars.iv, 16000 %arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3 ... and we must realize that the add will be a scalar in order to correctly deduce it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all dependencies of a consecutive pointer must be a scalar (uniform), and so we simply need to add all consecutive pointers to the worklist that currently detects collects uniforms. Fixes PR19296. llvm-svn: 205387
2014-04-02 10:34:49 +08:00
; CHECK: ret i32 0
}
attributes #0 = { nounwind }