public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug target/95532] New: Failure to optimize __builtin_ctz & 0x1F to __builtin_ctz on x86 with BMI
@ 2020-06-04 12:41 gabravier at gmail dot com
2020-06-04 15:54 ` [Bug target/95532] " jakub at gcc dot gnu.org
2020-06-04 15:57 ` gabravier at gmail dot com
0 siblings, 2 replies; 3+ messages in thread
From: gabravier at gmail dot com @ 2020-06-04 12:41 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95532
Bug ID: 95532
Summary: Failure to optimize __builtin_ctz & 0x1F to
__builtin_ctz on x86 with BMI
Product: gcc
Version: 11.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: gabravier at gmail dot com
Target Milestone: ---
int f(int x)
{
return __builtin_ctz(x) & 0x1F;
}
With -mbmi, this can be optimized to `__builtin_ctz(x)` (directly using `tzcnt`
without an `and` afterwards). This transformation is done by LLVM, but not by
GCC.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug target/95532] Failure to optimize __builtin_ctz & 0x1F to __builtin_ctz on x86 with BMI
2020-06-04 12:41 [Bug target/95532] New: Failure to optimize __builtin_ctz & 0x1F to __builtin_ctz on x86 with BMI gabravier at gmail dot com
@ 2020-06-04 15:54 ` jakub at gcc dot gnu.org
2020-06-04 15:57 ` gabravier at gmail dot com
1 sibling, 0 replies; 3+ messages in thread
From: jakub at gcc dot gnu.org @ 2020-06-04 15:54 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95532
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jakub at gcc dot gnu.org
--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Didn't you file the same for clz already?
The thing isthat at GIMPLE opts we include 32 in the range of possibly returned
values because it is a very common result for the 0 case and at RTL opts time
even know it will return 32 for 0, but the fact that it came up from a built
that was documented to have UB with zero value is lost there.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug target/95532] Failure to optimize __builtin_ctz & 0x1F to __builtin_ctz on x86 with BMI
2020-06-04 12:41 [Bug target/95532] New: Failure to optimize __builtin_ctz & 0x1F to __builtin_ctz on x86 with BMI gabravier at gmail dot com
2020-06-04 15:54 ` [Bug target/95532] " jakub at gcc dot gnu.org
@ 2020-06-04 15:57 ` gabravier at gmail dot com
1 sibling, 0 replies; 3+ messages in thread
From: gabravier at gmail dot com @ 2020-06-04 15:57 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95532
Gabriel Ravier <gabravier at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |DUPLICATE
Status|UNCONFIRMED |RESOLVED
--- Comment #2 from Gabriel Ravier <gabravier at gmail dot com> ---
Oops, now I remember doing that. smh my head, marking as duplicate of
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94801
*** This bug has been marked as a duplicate of bug 94801 ***
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-06-04 15:57 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-06-04 12:41 [Bug target/95532] New: Failure to optimize __builtin_ctz & 0x1F to __builtin_ctz on x86 with BMI gabravier at gmail dot com
2020-06-04 15:54 ` [Bug target/95532] " jakub at gcc dot gnu.org
2020-06-04 15:57 ` gabravier at gmail dot com
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).