https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
Richard Biener changed:
What|Removed |Added
Priority|P3 |P1
Status|NEW
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #20 from CVS Commits ---
The releases/gcc-13 branch has been updated by Jakub Jelinek
:
https://gcc.gnu.org/g:f980561c60b0446cc427595198d7f3f4f90e0924
commit r13-7231-gf980561c60b0446cc427595198d7f3f4f90e0924
Author: Andrew MacLeod
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #19 from CVS Commits ---
The master branch has been updated by Andrew Macleod :
https://gcc.gnu.org/g:17aa9ddb34581855dd013745c8be27dda024de4a
commit r14-122-g17aa9ddb34581855dd013745c8be27dda024de4a
Author: Andrew MacLeod
Date:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #18 from Andrew Macleod ---
(In reply to Richard Biener from comment #16)
> (In reply to Richard Biener from comment #15)
> > Created attachment 54892 [details]
> > patch I am testing
>
> Bootstrapped and tested on x86_64-unknown-li
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #17 from Andrew Macleod ---
Created attachment 54893
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54893&action=edit
new patch
Alternatively, we can simply allow undefined SSA_NAMES to also be checked or
single arguments like
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #16 from Richard Biener ---
(In reply to Richard Biener from comment #15)
> Created attachment 54892 [details]
> patch I am testing
Bootstrapped and tested on x86_64-unknown-linux-gnu, the XFAILs of
gcc.dg/pr102648.c and gcc.dg/pr10
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #15 from Richard Biener ---
Created attachment 54892
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54892&action=edit
patch I am testing
This is the patch I am testing which XFAILs the four testcases. It would be
nice to some
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #14 from Richard Biener ---
I think the issue is a bit more involved - when
_1 = PHI <_2, _3>
and _2 is UNDEFINED _on the edge!_, the equivalence _1 == _3 only holds
when the condition making _2 UNDEFINED holds. Trivially that's
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #13 from Richard Biener ---
(In reply to Martin Liška from comment #9)
> Simplified a bit more to:
>
> struct libkeccak_generalised_spec {
> int state_size;
> int word_size;
> } main_gspec;
>
> long gvar;
>
> int libkeccak_deg
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #12 from Richard Biener ---
Completely scrapping PHI equivalences will cause
FAIL: gcc.dg/pr102648.c scan-tree-dump-not optimized "foo"
FAIL: gcc.dg/pr103359.c scan-tree-dump-not evrp "c = 0"
FAIL: gcc.dg/tree-ssa/evrp-ignore.c scan
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
Jakub Jelinek changed:
What|Removed |Added
CC||jakub at gcc dot gnu.org
--- Comment #1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #10 from Richard Biener ---
(In reply to Richard Biener from comment #8)
> I fear something more fundamental is broken here and we
> possibly should play safe and not register any equivalence relation
> from PHIs.
>
> diff --git a/g
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #9 from Martin Liška ---
Simplified a bit more to:
struct libkeccak_generalised_spec {
int state_size;
int word_size;
} main_gspec;
long gvar;
int libkeccak_degeneralise_spec(struct libkeccak_generalised_spec *spec) {
int st
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #8 from Richard Biener ---
We basically see
if (!have_bitrate) {
state_size = (have_state_size ? state_size : (1600L));
output = ((state_size << 5) / 100L + 7L) & ~0x07L;
bitrate = output << 1;
}
optimized to
if (!have_b
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
Richard Biener changed:
What|Removed |Added
Attachment #54887|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #6 from Martin Liška ---
(In reply to Richard Biener from comment #4)
> Created attachment 54887 [details]
> testcase
Note the reduced test-case already started to fail with
r13-1938-g87dd4c8c83768a.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #5 from Richard Biener ---
So the failure mode is different than the previous case since there's no
backedge involved at all. Instead all of the VREL_EQ relations we add
are for PHIs like
output_73 = PHI
where we equate the resul
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #4 from Richard Biener ---
Created attachment 54887
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54887&action=edit
testcase
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
--- Comment #3 from Richard Biener ---
Created attachment 54886
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54886&action=edit
preprocessed source of the affected TU
it contains the single affected function libkeccak_degeneralise_spec a
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
Richard Biener changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109564
Richard Biener changed:
What|Removed |Added
Target Milestone|--- |13.0
Keywords|
21 matches
Mail list logo
