http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
Jakub Jelinek changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
Resolution|---
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #30 from vincenzo Innocente ---
better: as usual nastier bugs are in the tests!
[innocent@olsnba04 parallel]$ strace ./affinity-1.exe | & grep affin
execve("./affinity-1.exe", ["./affinity-1.exe"], [/* 61 vars */]) = 0
sched_getaffinit
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #29 from Jakub Jelinek ---
Created attachment 30967
--> http://gcc.gnu.org/bugzilla/attachment.cgi?id=30967&action=edit
Y
Ah, thanks, I can see where the failing sched_getaffinity calls are coming
from, hopefully this patch should f
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #28 from vincenzo Innocente ---
updated to the new revision
gcc version 4.9.0 20131007 (experimental) [gomp-4_0-branch revision 203250]
(GCC)
[innocent@olsnba04 parallel]$ setenv OMP_PROC_BIND 'spread,master,close';
setenv OMP_PLACES
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #27 from Jakub Jelinek ---
(In reply to vincenzo Innocente from comment #26)
> [innocent@olsnba04 parallel]$ ./affinity-1.exe | grep veri
>
> libgomp: Number of places reduced from 5 to 1 because some places didn't
> contain any usabl
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #26 from vincenzo Innocente ---
On 7 Oct, 2013, at 3:02 PM, jakub at gcc dot gnu.org
wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
>
> --- Comment #25 from Jakub Jelinek ---
> It is fine if the testcase doesn't fork an
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #25 from Jakub Jelinek ---
It is fine if the testcase doesn't fork and doesn't verify for taskset -c
24-31, that would be too hard to support, the testcase doesn't fail because of
that.
But, do you get the ", verified" strings in the l
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #24 from vincenzo Innocente ---
ok, modified to ">="
taskset -c 0-31 gdb ./affinity-1.exe
GNU gdb (GDB) Red Hat Enterprise Linux (7.2-60.el6_4.1)
(gdb) b
/home/data/newsoft/gcc-gomp4/libgomp/testsuite/libgomp.c/affinity-1.c:179
Breakp
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #23 from Jakub Jelinek ---
Ah, right, obviously I meant
--- libgomp/testsuite/libgomp.c/affinity-1.c2013-10-07 09:31:53.884695701
+0200
+++ libgomp/testsuite/libgomp.c/affinity-1.c2013-10-07 14:09:52.475331358
+0200
@@ -178,7 +
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #22 from vincenzo Innocente ---
on the XEON
setenv OMP_PROC_BIND false
reakpoint 1, main () at
/home/data/newsoft/gcc-gomp4/libgomp/testsuite/libgomp.c/affinity-1.c:181
181/home/data/newsoft/gcc-gomp4/libgomp/testsuite/libgomp.c/a
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #21 from Jakub Jelinek ---
Author: jakub
Date: Mon Oct 7 11:39:39 2013
New Revision: 203243
URL: http://gcc.gnu.org/viewcvs?rev=203243&root=gcc&view=rev
Log:
PR libgomp/58642
* config/linux/proc.c: Include errno.h.
(gomp_
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #20 from Jakub Jelinek ---
I'll commit the patch anyway, it is a step forward.
Anyway, the way the testcase is written is that if you run it with
OMP_PROC_BIND=false in the environment and OMP_PLACES unset, then it should
fork and exe
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #19 from vincenzo Innocente ---
On 7 Oct, 2013, at 12:27 PM, jakub at gcc dot gnu.org
wrote:
> or config.h doesn't defined HAVE_PTHREAD_AFFINITY_NP, then that's
> expected.
innocent@vinavx0 testsuite]$ cat ../config.h
/* config.h.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #18 from vincenzo Innocente ---
On 7 Oct, 2013, at 12:27 PM, jakub at gcc dot gnu.org
wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
>
> --- Comment #17 from Jakub Jelinek ---
> (In reply to vincenzo Innocente from comm
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #17 from Jakub Jelinek ---
(In reply to vincenzo Innocente from comment #16)
> [innocent@olsnba04 parallel]$ setenv OMP_PROC_BIND false
> [innocent@olsnba04 parallel]$ ./affinity-1.exe
> OMP_PROC_BIND='false'
> Initial thread
> #1 thre
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #16 from vincenzo Innocente ---
./affinity-1.exe
Initial thread
#1 thread 1
#1 thread 0
#1 thread 3
#1 thread 2
#1,#1 thread 3,1
#1,#1 thread 3,0
#1,#1 thread 3,2
#1,#2 thread 3,4
#1,#2 thread 3,0
#1,#2 thread 3,1
#1,#2 thread 3,3
#1
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #15 from Jakub Jelinek ---
(In reply to vincenzo Innocente from comment #14)
> for "historical reason" I build gcc on my local workstation not on the XEON.
> Also I do not have the gcc test infrastructure…
>
> can I copy any of those
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #14 from vincenzo Innocente ---
On 7 Oct, 2013, at 10:06 AM, jakub at gcc dot gnu.org
wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
>
> --- Comment #12 from Jakub Jelinek ---
> Ah, and can you also do make check RUNTES
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #13 from vincenzo Innocente ---
[innocent@olsnba04 parallel]$ setenv OMP_PROC_BIND true; setenv OMP_PLACES
'threads'
[innocent@olsnba04 parallel]$ gcc -fopenmp trivialOMP.cpp
[innocent@olsnba04 parallel]$ ./a.out
[innocent@olsnba04
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #12 from Jakub Jelinek ---
Ah, and can you also do make check RUNTESTFLAGS=c.exp=affinity-1.c
in the libgomp build directory and see whether it was PASS and, if
testsuite/libgomp.log contains any lines ending with verified and none mat
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #11 from Jakub Jelinek ---
(In reply to vincenzo Innocente from comment #10)
> seems working
Thanks. Can you please also try some simple (with -fopenmp):
int
main (void)
{
#pragma omp parallel num_threads (32)
asm volatile ("" :
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #10 from vincenzo Innocente ---
seems working
[innocent@olsnba04 parallel]$ c++ -std=c++11 -Ofast -fopenmp simpleOMP.cpp
[innocent@olsnba04 parallel]$ ./a.out
max thread 32
[innocent@olsnba04 parallel]$ taskset -c 0-3 ./a.out
max th
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
Jakub Jelinek changed:
What|Removed |Added
Status|UNCONFIRMED |ASSIGNED
Last reconfirmed|
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #8 from vincenzo Innocente ---
do you have access to a >32 cpu machine?
btw on XEON-PHI one can have >200 "cpus"
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #7 from vincenzo Innocente ---
getconf -a | grep _NPROCESSORS
_NPROCESSORS_CONF 32
_NPROCESSORS_ONLN 32
ls -l /sys/devices/system/cpu/
total 0
drwxr-xr-x 8 root root0 Sep 2 17:44 cpu0
drwxr-xr-x
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #6 from Jakub Jelinek ---
Also, can you dmesg | grep 'Allowing.*CPUs'
(or grep /var/log/messages* for the same)?
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #5 from Jakub Jelinek ---
So, looking at kernel sources, it seems it returns EINVAL:
if ((len * BITS_PER_BYTE) < nr_cpu_ids)
return -EINVAL;
so the question is how to find what your nr_cpu_ids is and how to (che
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #4 from vincenzo Innocente ---
24 thread machine ok
innocent@vocms19 parallel]$ c++ -Ofast -std=c++11 -fopenmp simpleOMP.cpp
[innocent@vocms19 parallel]$ ./a.out
max thread 24
[innocent@vocms19 parallel]$ taskset -c 0-3 ./a.out
max th
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #3 from vincenzo Innocente ---
strange indeed
rhel6: so is 2.12
or my own version
GNU C Library stable release version 2.13,
I build gcc by myself
c++ -v
Using built-in specs.
COLLECT_GCC=c++
COLLECT_LTO_WRAPPER=/afs/cern.ch/user/i/i
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #2 from Jakub Jelinek ---
What exact glibc do you use? I think the earliest that supports affinity in
libgomp should be 2.3, then between that and 2.7 there were multiple header
changes that the code just attempts to deal with using #
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58642
--- Comment #1 from Jakub Jelinek ---
Weird, can't reproduce.
/tmp/ttt
8
taskset -c 0-3 /tmp/ttt
4
ldd /tmp/ttt | grep libgomp
libgomp.so.1 =>
/usr/src/gomp-4.0/obj/x86_64-unknown-linux-gnu/libgomp/.libs/libgomp.so.1
(0x7fda5e2e8000)
strac
31 matches
Mail list logo