--- Comment #10 from sebastian dot pop at cri dot ensmp dot fr 2007-05-10
09:32 ---
Subject: Re: -ftree-vectorize results in internal compiler error on AMD64
Zdenek's patch for cleaning the dataref analysis is also fixing this bug.
http://gcc.gnu.org/ml/gcc-patches/2007-05/msg00634
--- Comment #1 from sebastian dot pop at cri dot ensmp dot fr 2006-11-06
11:48 ---
Subject: Re: Missed constant propagation into loops
I think the problem is that i is a global variable and thus foo is
potentially
considered as modifying i. Have you tried
void foo (void);
void bar
--- Comment #11 from sebastian dot pop at cri dot ensmp dot fr 2006-06-19
07:50 ---
Subject: Re: Number of iterations not know for simple loop
I thought that this bug should have been fixed by now:
http://gcc.gnu.org/ml/gcc-patches/2006-03/msg01749.html
what is the status
--- Comment #12 from sebastian dot pop at cri dot ensmp dot fr 2006-06-16
08:55 ---
Subject: Re: [4.2 Regression] segfault in fold_convert with -ftree-vectorize
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
The patch is in predcom branch.
I should have missed
--- Comment #9 from sebastian dot pop at cri dot ensmp dot fr 2006-06-15
13:19 ---
Subject: Re: [4.2 Regression] segfault in fold_convert with -ftree-vectorize
rakdver at gcc dot gnu dot org wrote:
Ccing people responsible for data dependence analysis.
Thanks, for the ping
--- Comment #10 from sebastian dot pop at cri dot ensmp dot fr 2006-06-15
14:52 ---
Subject: Re: [4.2 Regression] segfault in fold_convert with -ftree-vectorize
You said that you had a fix in predcom, is that fix in your local
tree, or have you sent a patch to gcc-patches?
Here
--- Comment #3 from sebastian dot pop at cri dot ensmp dot fr 2006-05-10
12:26 ---
Created an attachment (id=11429)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11429action=view)
fix
proposed fix. I didn't tested it other than making sure it fixed the bug.
--
http
--- Comment #5 from sebastian dot pop at cri dot ensmp dot fr 2006-05-10
12:32 ---
Created an attachment (id=11430)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11430action=view)
first shot at fixing this
didn't tested the patch, other than with RUNTESTFLAGS=tree-ssa.exp
--- Comment #12 from sebastian dot pop at cri dot ensmp dot fr 2006-04-10
09:14 ---
Created an attachment (id=11235)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11235action=view)
proposed fix
This patch fixes the problem, but probably it is a more general optimization
fix than
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sebastian dot pop at cri dot ensmp dot fr
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27063
--- Comment #12 from sebastian dot pop at cri dot ensmp dot fr 2006-04-02
08:12 ---
Created an attachment (id=11184)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11184action=view)
fix
This patch fixes this problem. I'll bootntestncommit.
--
http://gcc.gnu.org/bugzilla
--- Comment #3 from sebastian dot pop at cri dot ensmp dot fr 2006-03-29
20:39 ---
Fixed by the recent changes.
--
sebastian dot pop at cri dot ensmp dot fr changed:
What|Removed |Added
--- Comment #3 from sebastian dot pop at cri dot ensmp dot fr 2006-03-29
20:40 ---
Fixed on autovect.
--
sebastian dot pop at cri dot ensmp dot fr changed:
What|Removed |Added
--- Comment #2 from sebastian dot pop at cri dot ensmp dot fr 2006-03-29
20:42 ---
Fixed on autovect.
--
sebastian dot pop at cri dot ensmp dot fr changed:
What|Removed |Added
--- Comment #4 from sebastian dot pop at cri dot ensmp dot fr 2006-03-28
22:44 ---
Created an attachment (id=11146)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11146action=view)
first step
with this patch scev returns (int) (char) {0,+,1} but then
chrec_convert_aggressive
--- Comment #9 from sebastian dot pop at cri dot ensmp dot fr 2006-03-29
00:27 ---
Created an attachment (id=11147)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11147action=view)
proposed fix
this patch (not tested yet) fixes the problem: it avoids a division by zero.
Part
--- Comment #8 from sebastian dot pop at cri dot ensmp dot fr 2006-02-13
08:45 ---
Subject: Re: [4.0/4.1/4.2 Regression] DOM and VRP creating harder to optimize
code
This case reminds me the peeled chrec unification that I had to
disable on autovect branch (I probably have to run
--- Comment #4 from sebastian dot pop at cri dot ensmp dot fr 2006-01-05
16:32 ---
(In reply to comment #3)
Vectorizer's dump file can help.
vect dump ends with:
/home/seb/ex/pr25371.c:26: note: Access function of PHI: {0, +,
1}_4(get_loop_exit_condition
if (i_99 D.1700_228
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sebastian dot pop at cri dot ensmp dot fr
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24670
--- Comment #56 from sebastian dot pop at cri dot ensmp dot fr 2005-11-03
13:24 ---
Subject: Re: [4.0/4.1 Regression] IV-OPTS is O(N^3)
So, I'd suggest that we add a --param here for max-loop-nest-depth, and then
just not do this stuff on deeper nests, or ignore all the outer
--- Comment #57 from sebastian dot pop at cri dot ensmp dot fr 2005-11-03
15:02 ---
Subject: Re: [4.0/4.1 Regression] IV-OPTS is O(N^3)
sebastian dot pop at cri dot ensmp dot fr wrote:
So, I think that we can safely close this PR.
The previous numbers were with mainline + patch
--- Comment #58 from sebastian dot pop at cri dot ensmp dot fr 2005-11-03
19:31 ---
Subject: Re: [4.0/4.1 Regression] IV-OPTS is O(N^3)
Here are again the numbers for mainline with no other patch:
time ./gcc/cc1 -O2 ~/ex/pr18595_10.c
real0m0.164s
user0m0.116s
sys
--- Comment #59 from sebastian dot pop at cri dot ensmp dot fr 2005-11-03
20:10 ---
Subject: Re: [4.0/4.1 Regression] IV-OPTS is O(N^3)
Here is a first patch that uses PARAM_SCEV_MAX_EXPR_SIZE for limiting
the size of expressions that we want to handle. I will send it to
gcc-patches
--- Comment #4 from sebastian dot pop at cri dot ensmp dot fr 2005-10-12
16:26 ---
Subject: Re: [4.1 Regression] ICE: verify_ssa failed with -O -msse2
-ftree-vectorize
irar at il dot ibm dot com wrote:
Here scev analyzer calculates the evolution of 'D.1703_5 * 2 + i_15',
where 'D
--- Comment #5 from sebastian dot pop at cri dot ensmp dot fr 2005-10-12
16:53 ---
Subject: Re: [4.1 Regression] ICE: verify_ssa failed with -O -msse2
-ftree-vectorize
Sebastian Pop wrote:
if (!expr_invariant_in_loop_p (loop, CHREC_LEFT (chrec)))
then give up with this case
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-09-21 17:02 ---
Subject: Re: [4.1 Regression] loop problem / testcase takes very long time to
compile
Random break stops things typically somewhere inside 140 nested calls in scev
(follow_ssa_edge
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-09-14 12:38 ---
Subject: Re: [data deps] Distance on outer loops for self output deps
In this case neither implementation got the dependece right: there are
bugs in both implementations. For the following
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-22 12:34 ---
The following patch fixes the problem. However I cannot boot'n'regtest
it for the moment. I will commit it only tomorow once it is validated.
seb
*** tree-ssa-loop-niter.c.~2.39.~ 2005
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-22 12:38 ---
(In reply to comment #14)
(In reply to comment #9)
If we really wanted to tackle this better a compile-time, we'd run a
pass to look at all the ARRAY_REFs for those which have an out
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-18 11:51 ---
Reduced testcase is:
unsigned long long
f4 (unsigned long long diff)
{
return ((unsigned long long) ((signed long long) diff 0 ? -diff : diff));
}
main ()
{
int i;
for (i = 0; i = 10; i
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-17 23:23 ---
Subject: Re: [4.1 Regression] ICE: tree check: expected real_cst, have
integer_cst in const_binop, at fold-const.c:1512
I'm testing this patch on amd64 and i686. I will commit it once
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-15 12:33 ---
Subject: Re: [4.1 regression] Tree checking failure due to scev
pinskia at gcc dot gnu dot org wrote:
This is related to PR 19899 which was fixed.
Yes, PR is related to PR19899, but same
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-15 12:35 ---
Subject: Re: [4.1 regression] Tree checking failure due to scev
Sebastian Pop wrote:
Yes, PR is related to PR19899, but same pattern occured in several
places and the fix to PR19899
: normal
Priority: P2
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sebastian dot pop at cri dot ensmp dot fr
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23409
: 4.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sebastian dot pop at cri dot ensmp dot fr
CC: gcc-bugs at gcc dot gnu dot org
http
--
What|Removed |Added
BugsThisDependsOn||23411
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23409
--
What|Removed |Added
OtherBugsDependingO||23409
nThis||
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23411
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sebastian dot pop at cri dot ensmp dot fr
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla
--
What|Removed |Added
BugsThisDependsOn||23412
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23409
--
What|Removed |Added
OtherBugsDependingO||23409
nThis||
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23412
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sebastian dot pop at cri dot ensmp dot fr
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-15 23:16 ---
This one is also probably an overflow in Omega,
but the dependence problem looks pretty simple...
This occurs in gcc/gcc/omega.c on amd64-linux.
Dist vectors from the first dependence analyzer
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-08-15 00:34 ---
Subject: Re: [4.1 regression] Tree checking failure due to scev
This patch should fix the problem. There are some more cases that use
build_int_cst instead of build_real. I'll propose a more
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-07-27 09:01 ---
Subject: Re: [4.1 Regression] wrong code for casts and scev
dberlin at dberlin dot org wrote:
A sequence of unsigned char 1, 2, ..., 255 has to be converted to
signed char that would wrap
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-07-27 09:04 ---
Subject: Re: [4.1 Regression] wrong code for casts and scev
dberlin at dberlin dot org wrote:
I don't think it is possible to properly convert these ivs without
knowing an approximation
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-07-26 10:06 ---
Subject: Re: [4.1 Regression] wrong code for casts and scev
After inlining, we end up with a loop containing the following code:
b.0_3 = (signed char) b_8;
D.1621_4 = (int) b.0_3;
a_5
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-07-26 15:15 ---
Subject: Re: [4.1 Regression] wrong code for casts and scev
Dorit Naishlos wrote:
The modifications you suggest will make the tests uninteresting - they were
introduced with unknown loop
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-06-07 20:02 ---
fixed.
--
What|Removed |Added
Status|NEW
Severity: normal
Priority: P2
Component: tree-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sebastian dot pop at cri dot ensmp dot fr
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21861
--
What|Removed |Added
BugsThisDependsOn||18403, 21029
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21861
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-04-29 19:51 ---
Subject: Re: [PR tree-optimization/21029, RFC] harmful chrec type conversions
Thanks to Roger Sayle for pointing me at this PR.
Alexandre Oliva wrote:
This is not a final patch; if the idea
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-04-04 14:50 ---
Subject: Re: [4.0/4.1 Regression] Hang in tree_ssa_iv_optimize_loop
rakdver at gcc dot gnu dot org wrote:
Scev probably should keep track of how large expressions it produces, and
just
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-27 13:18 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
With the following patch I got some speedup for depth 100.
from:
tree iv optimization : 2.62 (62%) usr 0.27 (82%) sys 2.92 (62%) wall
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-27 15:12 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
the patch is below (in stronger form -- only removing entries that
contain
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-27 20:38 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
Another idea: would it be possible to insert the invalidated names
during
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-25 10:32 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
Adding the instantiation cache was compile time neutral on normal code,
so I don't
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-25 10:39 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
How? If the reference is left in symbolic form, it means that you know
nothing
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-25 11:02 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
(*) I hope; scev is a mess of mutualy recursive functions
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-25 12:03 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
--- Additional Comments From rakdver at atrey dot karlin dot mff dot cuni
dot
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-25 12:44 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
rakdver at atrey dot karlin dot mff dot cuni dot cz wrote:
More seriously -- which of the possibilities? If I have loops like
while
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-25 16:26 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
If you decide to reconstruct the tree expression, there is no reason
to stop on a phi node that has a single argument. Why would you like
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-24 21:36 ---
Subject: Re: [4.0 Regression] IV-OPTS is O(N^3)
Still there remain some inefficiences within the scev analysis itself.
Zdenek, have you tried to revert the patch that caches
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-02 20:37 ---
(In reply to comment #5)
Caused by an exponential time complexity of instantiate_parameters.
I am working on a patch.
The following patch solves the exponential time complexity.
Sorry
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-02 21:46 ---
(In reply to comment #7)
Subject: Re: [4.0 regression] Endless loop compiling simple file: Bug in
tree-scalar-evolution.c (instantiate_parameters_1)?
not really (well, maybe
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2005-01-02 23:15 ---
Subject: Re: [4.0 regression] Endless loop compiling simple file: Bug in
tree-scalar-evolution.c (instantiate_parameters_1)?
rakdver at gcc dot gnu dot org wrote:
--- Additional
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr
2004-12-19 23:43 ---
Patch in autovect-branch:
http://gcc.gnu.org/ml/gcc-patches/2004-12/msg01444.html
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17100
--- Additional Comments From sebastian dot pop at cri dot ensmp dot fr 2004-10-19
10:03 ---
Subject: Re: [4.0 Regression] ICE with simple loop with VLA
Patch is here:
http://gcc.gnu.org/ml/gcc-patches/2004-10/msg01592.html
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18005
67 matches
Mail list logo