[llvm] [LV][VPlan] Remove any-of reduction from precomputeCost. NFC (PR #117109)

LLVM Continuous Integration via llvm-commits llvm-commits at lists.llvm.org
Thu Nov 21 20:56:25 PST 2024


llvm-ci wrote:

LLVM Buildbot has detected a new failure on builder `lld-x86_64-ubuntu-fast` running on `as-builder-4` while building `llvm` at step 6 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/33/builds/6980

<details>
<summary>Here is the relevant piece of the build log for the reference</summary>

```
Step 6 (test-build-unified-tree-check-all) failure: test (failure)
******************** TEST 'LLVM :: Transforms/LoopVectorize/AArch64/select-costs.ll' FAILED ********************
Exit Code: 1

Command Output (stderr):
--
RUN: at line 2: /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/build/bin/opt < /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/llvm-project/llvm/test/Transforms/LoopVectorize/AArch64/select-costs.ll -passes=loop-vectorize -debug-only=loop-vectorize -disable-output -S 2>&1 | /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/build/bin/FileCheck /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/llvm-project/llvm/test/Transforms/LoopVectorize/AArch64/select-costs.ll
+ /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/build/bin/opt -passes=loop-vectorize -debug-only=loop-vectorize -disable-output -S
+ /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/build/bin/FileCheck /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/llvm-project/llvm/test/Transforms/LoopVectorize/AArch64/select-costs.ll
/home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/llvm-project/llvm/test/Transforms/LoopVectorize/AArch64/select-costs.ll:54:10: error: CHECK: expected string not found in input
; CHECK: any-of reduction %all.off = select i1 %cmp1, i1 %all.off.next, i1 false
         ^
<stdin>:806:21: note: scanning from here
Cost of 1 for VF 16: induction instruction %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
                    ^
<stdin>:832:47: note: possible intended match here
LV: Found an estimated cost of 1 for VF 1 For instruction: %all.off = select i1 %cmp1, i1 %all.off.next, i1 false
                                              ^

Input file: <stdin>
Check file: /home/buildbot/worker/as-builder-4/ramdisk/lld-x86_64/llvm-project/llvm/test/Transforms/LoopVectorize/AArch64/select-costs.ll

-dump-input=help explains the following input dump.

Input was:
<<<<<<
            .
            .
            .
          801: Cost of 0 for VF 8: EMIT vp<%7> = or ir<%all.off.next>, vp<%6> 
          802: Cost of 0 for VF 8: EMIT vp<%index.next> = add nuw vp<%2>, vp<%0> 
          803: Cost of 0 for VF 8: EMIT branch-on-count vp<%index.next>, vp<%1> 
          804: Cost of 0 for VF 8: vector loop backedge 
          805: Cost for VF 8: 6 (Estimated cost per lane: 0.8) 
          806: Cost of 1 for VF 16: induction instruction %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1 
check:54'0                         X~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ error: no match found
          807: Cost of 0 for VF 16: induction instruction %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] 
check:54'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
          808: Cost of 1 for VF 16: exit condition instruction %exitcond.not = icmp eq i64 %indvars.iv.next, %n 
check:54'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
          809: Cost of 0 for VF 16: forced scalar %arrayidx = getelementptr inbounds float, ptr %a, i64 %indvars.iv 
check:54'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
          810: Cost of 0 for VF 16: EMIT vp<%2> = CANONICAL-INDUCTION ir<0>, vp<%index.next> 
check:54'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
          811: Cost of 0 for VF 16: WIDEN-REDUCTION-PHI ir<%all.off.next> = phi ir<false>, vp<%7> 
check:54'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
            .
            .
            .
          827: LV: Found an estimated cost of 0 for VF 1 For instruction: %any.0.off09 = phi i1 [ false, %entry ], [ %.any.0.off0, %for.body ] 
check:54'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
...

```

</details>

https://github.com/llvm/llvm-project/pull/117109


More information about the llvm-commits mailing list