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/sink-out-of-loop.ll

76 lines
2.0 KiB
LLVM
Raw Normal View History

MachineSink: Sort the successors before trying to find SuccToSinkTo. One motivating example is to sink an instruction from a basic block which has two successors: one outside the loop, the other inside the loop. We should try to sink the instruction outside the loop. rdar://11980766 llvm-svn: 161062
2012-08-01 02:10:39 +08:00
; RUN: llc -mtriple=x86_64-apple-darwin < %s | FileCheck %s
; A MOV32ri is inside a loop, it has two successors, one successor is inside the
; same loop, the other successor is outside the loop. We should be able to sink
; MOV32ri outside the loop.
; rdar://11980766
define i32 @sink_succ(i32 %argc, i8** nocapture %argv) nounwind uwtable ssp {
[MachineSink] Use the real post dominator tree Summary: Fixes a FIXME in MachineSinking. Instead of using the simple heuristics in isPostDominatedBy, use the real MachinePostDominatorTree and MachineLoopInfo. The old heuristics caused instructions to sink unnecessarily, and might create register pressure. This is the second try of the fix. The first one (D4814) caused a performance regression due to failing to sink instructions out of loops (PR21115). This patch fixes PR21115 by sinking an instruction from a deeper loop to a shallower one regardless of whether the target block post-dominates the source. Thanks Alexey Volkov for reporting PR21115! Test Plan: Added a NVPTX codegen test to verify that our change prevents the backend from over-sinking. It also shows the unnecessary register pressure caused by over-sinking. Added an X86 test to verify we can sink instructions out of loops regardless of the dominance relationship. This test is reduced from Alexey's test in PR21115. Updated an affected test in X86. Also ran SPEC CINT2006 and llvm-test-suite for compilation time and runtime performance. Results are attached separately in the review thread. Reviewers: Jiangning, resistor, hfinkel Reviewed By: hfinkel Subscribers: hfinkel, bruno, volkalexey, llvm-commits, meheff, eliben, jholewinski Differential Revision: http://reviews.llvm.org/D5633 llvm-svn: 219773
2014-10-15 11:27:43 +08:00
; CHECK-LABEL: sink_succ
MachineSink: Sort the successors before trying to find SuccToSinkTo. One motivating example is to sink an instruction from a basic block which has two successors: one outside the loop, the other inside the loop. We should try to sink the instruction outside the loop. rdar://11980766 llvm-svn: 161062
2012-08-01 02:10:39 +08:00
; CHECK: [[OUTER_LN1:LBB0_[0-9]+]]: ## %preheader
; CHECK: %exit
; CHECK-NOT: movl
; CHECK: jne [[OUTER_LN1]]
; CHECK: movl
; CHECK: [[LN2:LBB0_[0-9]+]]: ## %for.body2
; CHECK: jne [[LN2]]
; CHECK: ret
entry:
br label %preheader
preheader:
%i.127 = phi i32 [ 0, %entry ], [ %inc9, %exit ]
br label %for.body1.lr
for.body1.lr:
%iv30 = phi i32 [ 1, %preheader ], [ %iv.next31, %for.inc40.i ]
br label %for.body1
for.body1:
%iv.i = phi i64 [ 0, %for.body1.lr ], [ %iv.next.i, %for.body1 ]
%iv.next.i = add i64 %iv.i, 1
%lftr.wideiv32 = trunc i64 %iv.next.i to i32
%exitcond33 = icmp eq i32 %lftr.wideiv32, %iv30
br i1 %exitcond33, label %for.inc40.i, label %for.body1
for.inc40.i:
%iv.next31 = add i32 %iv30, 1
%exitcond49.i = icmp eq i32 %iv.next31, 32
br i1 %exitcond49.i, label %exit, label %for.body1.lr
exit:
%inc9 = add nsw i32 %i.127, 1
%exitcond34 = icmp eq i32 %inc9, 10
br i1 %exitcond34, label %for.body2, label %preheader
for.body2:
%iv = phi i64 [ %iv.next, %for.body2 ], [ 0, %exit ]
%iv.next = add i64 %iv, 1
%lftr.wideiv = trunc i64 %iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, 2048
br i1 %exitcond, label %for.end20, label %for.body2
for.end20:
ret i32 0
}
[MachineSink] Use the real post dominator tree Summary: Fixes a FIXME in MachineSinking. Instead of using the simple heuristics in isPostDominatedBy, use the real MachinePostDominatorTree and MachineLoopInfo. The old heuristics caused instructions to sink unnecessarily, and might create register pressure. This is the second try of the fix. The first one (D4814) caused a performance regression due to failing to sink instructions out of loops (PR21115). This patch fixes PR21115 by sinking an instruction from a deeper loop to a shallower one regardless of whether the target block post-dominates the source. Thanks Alexey Volkov for reporting PR21115! Test Plan: Added a NVPTX codegen test to verify that our change prevents the backend from over-sinking. It also shows the unnecessary register pressure caused by over-sinking. Added an X86 test to verify we can sink instructions out of loops regardless of the dominance relationship. This test is reduced from Alexey's test in PR21115. Updated an affected test in X86. Also ran SPEC CINT2006 and llvm-test-suite for compilation time and runtime performance. Results are attached separately in the review thread. Reviewers: Jiangning, resistor, hfinkel Reviewed By: hfinkel Subscribers: hfinkel, bruno, volkalexey, llvm-commits, meheff, eliben, jholewinski Differential Revision: http://reviews.llvm.org/D5633 llvm-svn: 219773
2014-10-15 11:27:43 +08:00
define i32 @sink_out_of_loop(i32 %n, i32* %output) {
; CHECK-LABEL: sink_out_of_loop:
entry:
br label %loop
loop:
%i = phi i32 [ 0, %entry ], [ %i2, %loop ]
%j = mul i32 %i, %i
[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
%addr = getelementptr i32, i32* %output, i32 %i
[MachineSink] Use the real post dominator tree Summary: Fixes a FIXME in MachineSinking. Instead of using the simple heuristics in isPostDominatedBy, use the real MachinePostDominatorTree and MachineLoopInfo. The old heuristics caused instructions to sink unnecessarily, and might create register pressure. This is the second try of the fix. The first one (D4814) caused a performance regression due to failing to sink instructions out of loops (PR21115). This patch fixes PR21115 by sinking an instruction from a deeper loop to a shallower one regardless of whether the target block post-dominates the source. Thanks Alexey Volkov for reporting PR21115! Test Plan: Added a NVPTX codegen test to verify that our change prevents the backend from over-sinking. It also shows the unnecessary register pressure caused by over-sinking. Added an X86 test to verify we can sink instructions out of loops regardless of the dominance relationship. This test is reduced from Alexey's test in PR21115. Updated an affected test in X86. Also ran SPEC CINT2006 and llvm-test-suite for compilation time and runtime performance. Results are attached separately in the review thread. Reviewers: Jiangning, resistor, hfinkel Reviewed By: hfinkel Subscribers: hfinkel, bruno, volkalexey, llvm-commits, meheff, eliben, jholewinski Differential Revision: http://reviews.llvm.org/D5633 llvm-svn: 219773
2014-10-15 11:27:43 +08:00
store i32 %i, i32* %addr
%i2 = add i32 %i, 1
%exit_cond = icmp sge i32 %i2, %n
br i1 %exit_cond, label %exit, label %loop
exit:
; CHECK: BB#2
; CHECK: imull %eax, %eax
; CHECK: retq
ret i32 %j
}