This is an archive of the discontinued LLVM Phabricator instance.

Differential D151617

A test case for the revision D151616
AbandonedPublic

Authored by pawosm01 on May 27 2023, 11:57 AM.

Details

Reviewers
qcolombet
Summary

This commit introduces a test for the change addressing the following
issue: https://github.com/llvm/llvm-project/issues/62736

The change commit itself should also remove the FIXME comment.

Diff Detail

Event Timeline

pawosm01 created this revision.May 27 2023, 11:57 AM
Herald added a project: Restricted Project. · View Herald TranscriptMay 27 2023, 11:57 AM
pawosm01 requested review of this revision.May 27 2023, 11:57 AM
Herald added a project: Restricted Project. · View Herald TranscriptMay 27 2023, 11:57 AM
Herald added a subscriber: llvm-commits. · View Herald Transcript
pawosm01 added a child revision: D151616: [Transforms][Reassociate] "Reassociate expressions" pass optimizations not always profitable.May 27 2023, 12:00 PM
Harbormaster completed remote builds in B235030: Diff 526281.May 27 2023, 12:52 PM
qcolombet accepted this revision.May 31 2023, 4:41 AM
qcolombet added inline comments.
llvm/test/Transforms/Reassociate/reassociate-not-from-the-outside-of-the-loop.ll
1

Maybe use update_test_checks.py here as well so that the diff will be obvious in your next PR.

30

Please run instnamer to get rid of the implicit variables.

This revision is now accepted and ready to land.May 31 2023, 4:41 AM
pawosm01 abandoned this revision.EditedJun 7 2023, 9:18 AM

Overriden by D152282

Herald added a subscriber: mgabka. · View Herald TranscriptJun 7 2023, 9:18 AM

Revision Contents

PathSize
llvm/
test/
Transforms/
Reassociate/
44 lines

Diff 526281

llvm/test/Transforms/Reassociate/reassociate-not-from-the-outside-of-the-loop.ll

  • This file was added.
; RUN: opt -passes=reassociate -S < %s | FileCheck %s
qcolombetUnsubmitted
Not Done
ReplyInline Actions

Maybe use update_test_checks.py here as well so that the diff will be obvious in your next PR.

qcolombet: Maybe use `update_test_checks.py` here as well so that the diff will be obvious in your next PR.
; This test is to ensure that no computations are pulled into a loop
; by the Reassociate pass. Doing so can result in the loop invariants not being
; computed before the loop anymore. In case of this test, it would add an extra
; multiplication into the loop.
; FIXME: the checks below need to be inverted to confirm the change to the
; Reassociate pass.
define void @innermost_loop(i32 %i, double %d1, double %d2, double %delta, ptr %cells) {
; CHECK-LABEL: @innermost_loop(
entry:
; CHECK-LABEL: entry:
%mul = fmul fast double %d1, %delta
%mul1 = fmul fast double %d2, %delta
; CHECK-NOT: %{{.*}} = fmul {{.*}} %delta
br label %for.cond
for.cond:
%j.0 = phi i32 [ 0, %entry ], [ %add, %for.body ]
%cmp.not = icmp sgt i32 %j.0, %i
br i1 %cmp.not, label %for.end, label %for.body
for.body:
; CHECK-LABEL: for.body:
%add = add nuw nsw i32 %j.0, 1
%idxprom = zext i32 %add to i64
%arrayidx = getelementptr inbounds double, ptr %cells, i64 %idxprom
%0 = load double, ptr %arrayidx
qcolombetUnsubmitted
Not Done
ReplyInline Actions

Please run instnamer to get rid of the implicit variables.

qcolombet: Please run `instnamer` to get rid of the implicit variables.
%mul2 = fmul fast double %mul, %0
%idxprom3 = zext i32 %j.0 to i64
%arrayidx4 = getelementptr inbounds double, ptr %cells, i64 %idxprom3
%1 = load double, ptr %arrayidx4
%mul5 = fmul fast double %mul1, %1
%add6 = fadd fast double %mul2, %mul5
; CHECK: %reass{{.*}} = fadd
; CHECK-NEXT: %reass{{.*}} = fmul {{.*}} %delta
store double %add6, ptr %arrayidx4
br label %for.cond
for.end:
ret void
}