public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug c++/115018] New: Incorrect data read/written at -O3 on Arm with 256-bit SVE SIMD width specified
@ 2024-05-09 20:29 erik at kth dot se
2024-05-09 21:05 ` [Bug target/115018] Incorrect data read/written at -O3 on aarch64 " pinskia at gcc dot gnu.org
2024-05-09 21:11 ` pinskia at gcc dot gnu.org
0 siblings, 2 replies; 3+ messages in thread
From: erik at kth dot se @ 2024-05-09 20:29 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115018
Bug ID: 115018
Summary: Incorrect data read/written at -O3 on Arm with 256-bit
SVE SIMD width specified
Product: gcc
Version: 13.2.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: erik at kth dot se
Target Milestone: ---
Created attachment 58154
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=58154&action=edit
Source code to reproduce the bug (only includes 3x std headers)
Hi GCC,
When testing Gromacs on Amazon's Graviton3 Arm nodes with SVE support using
Ubuntu 24.04 with gcc-13.2.0, I have been tracking down a strange failed unit
test that occurs in non-SVE code, but only when I manually specify 256-bit SVE
length (instead of auto).
This code had a ton of dependencies on Google test and other files, but I think
I have been able to isolated it to a small file that only includes vector and
array - see attachment.
The code in question creates a std::vector with 10 elements, where each element
is a std::array with 3 floats. Initially they are all set to zero, and then I
add (0,10,0) to all elements.
With this combination of high optimization and 256-bit SVE, only some of the
elements will be updated, while the other ones are still (0,0,0):
g++ -O3 -march=armv8.2-a+sve -msve-vector-bits=256 bug.cpp -Wall -Wextra
-fno-strict-aliasing -fwrapv -fno-aggressive-loop-optimizations
If I remove either then -O3, or -msve-vector-bits=256, the results are fine.
The bug also disappears if I don't have this slightly-complex multi-level
inheritance, or if I use plain C arrays instead of std containers, or remove
any of the other seemingly stupid constructs in the attached example - but at
least I saved you from having to debug all of Google test :-)
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug target/115018] Incorrect data read/written at -O3 on aarch64 with 256-bit SVE SIMD width specified
2024-05-09 20:29 [Bug c++/115018] New: Incorrect data read/written at -O3 on Arm with 256-bit SVE SIMD width specified erik at kth dot se
@ 2024-05-09 21:05 ` pinskia at gcc dot gnu.org
2024-05-09 21:11 ` pinskia at gcc dot gnu.org
1 sibling, 0 replies; 3+ messages in thread
From: pinskia at gcc dot gnu.org @ 2024-05-09 21:05 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115018
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
First I can't reproduce it with the official releases GCC 13.2.0.
Second with the trunk, I am not sure if this is a bug or not.
With the trunk with `-O3 -Wall -Wextra -fno-strict-aliasing -fwrapv
-fno-aggressive-loop-optimizations -march=armv8.2-a+sve -msve-vector-bits=256 `
Running under qemu like this:
[apinski@xeond2 upstream-cross-aarch64]$ ./install-qemu/bin/qemu-aarch64 -cpu
max,sve128=on,sve256=on a.out
x[ 0]: 0 0 0
x[ 1]: 0 0 0
x[ 2]: 0 0 0
x[ 3]: 0 0 0
x[ 4]: 0 0 0
x[ 5]: 0 0 0
x[ 6]: 0 0 0
x[ 7]: 0 0 0
x[ 8]: 0 0 0
x[ 9]: 0 0 0
x[ 0]: 0 10 0
x[ 1]: 0 10 0
x[ 2]: 0 10 0
x[ 3]: 0 10 0
x[ 4]: 0 10 0
x[ 5]: 0 10 0
x[ 6]: 0 10 0
x[ 7]: 0 10 0
x[ 8]: 0 10 0
x[ 9]: 0 10 0
Works.
But if I enable SVE512, it fails like you are describing:
[apinski@xeond2 upstream-cross-aarch64]$ ./install-qemu/bin/qemu-aarch64 -cpu
max,sve128=on,sve256=on,sve512=on a.out
x[ 0]: 0 0 0
x[ 1]: 0 0 0
x[ 2]: 0 0 0
x[ 3]: 0 0 0
x[ 4]: 0 0 0
x[ 5]: 0 0 0
x[ 6]: 0 0 0
x[ 7]: 0 0 0
x[ 8]: 0 0 0
x[ 9]: 0 0 0
x[ 0]: 0 10 0
x[ 1]: 0 10 0
x[ 2]: 0 10 0
x[ 3]: 0 0 0
x[ 4]: 0 0 0
x[ 5]: 0 10 0
x[ 6]: 0 10 0
x[ 7]: 0 10 0
x[ 8]: 0 0 0
x[ 9]: 0 0 0
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug target/115018] Incorrect data read/written at -O3 on aarch64 with 256-bit SVE SIMD width specified
2024-05-09 20:29 [Bug c++/115018] New: Incorrect data read/written at -O3 on Arm with 256-bit SVE SIMD width specified erik at kth dot se
2024-05-09 21:05 ` [Bug target/115018] Incorrect data read/written at -O3 on aarch64 " pinskia at gcc dot gnu.org
@ 2024-05-09 21:11 ` pinskia at gcc dot gnu.org
1 sibling, 0 replies; 3+ messages in thread
From: pinskia at gcc dot gnu.org @ 2024-05-09 21:11 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115018
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2024-05-09
Ever confirmed|0 |1
Status|UNCONFIRMED |WAITING
--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
https://community.arm.com/support-forums/f/high-performance-computing-forum/54025/size-of-vectors-in-sve2
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2024-05-09 21:11 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-05-09 20:29 [Bug c++/115018] New: Incorrect data read/written at -O3 on Arm with 256-bit SVE SIMD width specified erik at kth dot se
2024-05-09 21:05 ` [Bug target/115018] Incorrect data read/written at -O3 on aarch64 " pinskia at gcc dot gnu.org
2024-05-09 21:11 ` pinskia at gcc dot gnu.org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).