Re: [r16-5526 Regression] FAIL: gcc.dg/vect/bb-slp-41.c scan-tree-dump-not slp1 "vectorizing stmts using SLP" on Linux/x86_64

Hi,

I can reproduce this on x86_64. The issue is that bb-slp-41.c expects BB vevtorization to not happen. Look like the original test case had other checks too that has since gone.
I have added -fno-tree-loop-im and tested to keep the behaviour the same. Is this OK?

Thanks,
Kugan



> On 23 Nov 2025, at 7:20 pm, Haochen Jiang <[email protected]> wrote:
>
> External email: Use caution opening links or attachments
>
>
> On Linux/x86_64,
>
> ed1911b9f1f1ab0c1b631f0b6427b798c7056200 is the first bad commit
> commit ed1911b9f1f1ab0c1b631f0b6427b798c7056200
> Author: Kugan Vivekanandarajah <[email protected]>
> Date:   Sun Nov 23 15:27:10 2025 +1100
>
>    [tree-optimization] Allow LICM to hoist loads in "self write" patterns
>
> caused
>
> FAIL: gcc.dg/vect/bb-slp-41.c -flto -ffat-lto-objects  scan-tree-dump-not slp1 "vectorizing stmts using SLP"
> FAIL: gcc.dg/vect/bb-slp-41.c scan-tree-dump-not slp1 "vectorizing stmts using SLP"
>
> with GCC configured with
>
> ../../gcc/configure --prefix=/export/users3/haochenj/src/gcc-bisect/master/master/r16-5526/usr --enable-clocale=gnu --with-system-zlib --with-demangler-in-ld --with-fpmath=sse --enable-languages=c,c++,fortran --enable-cet --without-isl --enable-libmpx x86_64-linux --disable-bootstrap
>
> To reproduce:
>
> $ cd {build_dir}/gcc && make check RUNTESTFLAGS="vect.exp=gcc.dg/vect/bb-slp-41.c --target_board='unix{-m32}'"
> $ cd {build_dir}/gcc && make check RUNTESTFLAGS="vect.exp=gcc.dg/vect/bb-slp-41.c --target_board='unix{-m32\ -march=cascadelake}'"
> $ cd {build_dir}/gcc && make check RUNTESTFLAGS="vect.exp=gcc.dg/vect/bb-slp-41.c --target_board='unix{-m64\ -march=cascadelake}'"
>
> (Please directly reply to this email for question about this report.)
> (If you met problems with cascadelake related, disabling AVX512F in command line might save that.)
> (However, please make sure that there is no potential problems with AVX512.)

From 6a37cb2081eec60d4546e690bdf02ebe3baa483a Mon Sep 17 00:00:00 2001
From: Kugan Vivekanandarajah <[email protected]>
Date: Mon, 24 Nov 2025 11:08:44 +1100
Subject: [PATCH] [TESTSUITE][bb-slp-41.c] Adjust to prevent LICM to keep the
 expected behaviour

Now that we do move b[0] in LICM, this is
changing the expected behaviour in:
  for (i = 0; i < (ARR_SIZE - 2); ++i)
  {
    a[i] = b[0];
  }

Added -fno-tree-loop-im to preseve the exisitng behaviour.

gcc/testsuite/ChangeLog:

2025-11-24  Kugan Vivekanandarajah  <[email protected]>

	* gcc.dg/vect/bb-slp-41.c: Adjust to prevent loop invariant code
	motion to keep the expected behaviour.

Signed-off-by: Kugan Vivekanandarajah <[email protected]>
---
 gcc/testsuite/gcc.dg/vect/bb-slp-41.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gcc/testsuite/gcc.dg/vect/bb-slp-41.c b/gcc/testsuite/gcc.dg/vect/bb-slp-41.c
index 5a2bd4d2a33..d7937fa7148 100644
--- a/gcc/testsuite/gcc.dg/vect/bb-slp-41.c
+++ b/gcc/testsuite/gcc.dg/vect/bb-slp-41.c
@@ -1,4 +1,5 @@
 /* { dg-require-effective-target vect_int } */
+/* { dg-additional-options "-fno-tree-loop-im" } */
 
 #define ARR_SIZE 1000
 
-- 
2.50.1 (Apple Git-155)