https://bugzilla.redhat.com/show_bug.cgi?id=739398
Fedora Update System changed:
What|Removed |Added
Fixed In Version|openblas-0.2.5-10.el6 |openblas-0.2.11-1.el7
--- Comme
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #51 from Fedora Update System ---
openblas-0.2.11-1.el7 has been submitted as an update for Fedora EPEL 7.
https://admin.fedoraproject.org/updates/openblas-0.2.11-1.el7
--
You are receiving this mail because:
You are on the CC lis
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #50 from Jon Ciesla ---
Git done (by process-git-requests).
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
__
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Jon Ciesla changed:
What|Removed |Added
Flags|fedora-cvs? |fedora-cvs+
--
You are receiving this ma
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Susi Lehtola changed:
What|Removed |Added
Flags|fedora-cvs+ |fedora-cvs?
--- Comment #49 from Susi L
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Dominik 'Rathann' Mierzejewski changed:
What|Removed |Added
Blocks|505154 (FE-SCITECH) |
Referenced Bugs:
h
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #48 from Susi Lehtola ---
No, I hadn't thought about it at all...
I believe we can't make the 64-bit interface the default, because this needs to
be taken care of in the user side as well.
Anyway, patches are welcome.
Also, this
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Milan Bouchet-Valat changed:
What|Removed |Added
CC||nalimi...@club.fr
--- Comment #4
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Fedora Update System changed:
What|Removed |Added
Fixed In Version|openblas-0.2.5-10.fc19 |openblas-0.2.5-10.el5
--- Comment
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Fedora Update System changed:
What|Removed |Added
Fixed In Version|openblas-0.2.5-10.el5 |openblas-0.2.5-10.el6
--- Comment
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Fedora Update System changed:
What|Removed |Added
Fixed In Version|openblas-0.2.5-10.fc18 |openblas-0.2.5-10.fc19
--- Commen
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Fedora Update System changed:
What|Removed |Added
Status|ON_QA |CLOSED
Fixed In Version|
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #42 from Susi Lehtola ---
FYI, this was hanging around since I was waiting for upstream to fix a rather
nasty bug [1], and only recently received information that the bug is in the
OpenBLAS LAPACK functions. I've now replaced the opti
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #41 from Fedora Update System ---
openblas-0.2.5-10.el5 has been submitted as an update for Fedora EPEL 5.
https://admin.fedoraproject.org/updates/openblas-0.2.5-10.el5
--
You are receiving this mail because:
You are on the CC list
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #40 from Fedora Update System ---
openblas-0.2.5-10.fc18 has been submitted as an update for Fedora 18.
https://admin.fedoraproject.org/updates/openblas-0.2.5-10.fc18
--
You are receiving this mail because:
You are on the CC list fo
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #39 from Fedora Update System ---
openblas-0.2.5-10.el6 has been submitted as an update for Fedora EPEL 6.
https://admin.fedoraproject.org/updates/openblas-0.2.5-10.el6
--
You are receiving this mail because:
You are on the CC list
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #38 from Fedora Update System ---
openblas-0.2.5-10.fc19 has been submitted as an update for Fedora 19.
https://admin.fedoraproject.org/updates/openblas-0.2.5-10.fc19
--
You are receiving this mail because:
You are on the CC list fo
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Fedora Update System changed:
What|Removed |Added
Status|MODIFIED|ON_QA
--
You are
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #37 from Fedora Update System ---
openblas-0.2.5-7.el5 has been pushed to the Fedora EPEL 5 testing repository.
--
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug
h
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #36 from Fedora Update System ---
openblas-0.2.5-7.fc17 has been submitted as an update for Fedora 17.
https://admin.fedoraproject.org/updates/openblas-0.2.5-7.fc17
--
You are receiving this mail because:
You are on
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #35 from Fedora Update System ---
openblas-0.2.5-7.el5 has been submitted as an update for Fedora EPEL 5.
https://admin.fedoraproject.org/updates/openblas-0.2.5-7.el5
--
You are receiving this mail because:
You are o
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #34 from Fedora Update System ---
openblas-0.2.5-7.fc18 has been submitted as an update for Fedora 18.
https://admin.fedoraproject.org/updates/openblas-0.2.5-7.fc18
--
You are receiving this mail because:
You are on
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #33 from Fedora Update System ---
openblas-0.2.5-7.el6 has been submitted as an update for Fedora EPEL 6.
https://admin.fedoraproject.org/updates/openblas-0.2.5-7.el6
--
You are receiving this mail because:
You are o
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Fedora Update System changed:
What|Removed |Added
Status|ASSIGNED|MODIFIED
--
You
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #32 from Susi Lehtola ---
For reference, I'm not building any non-x86(_64) binaries since upstream only
supports x86 and Loongson; they don't have access to ia64, ppc, sparc and power
architecture machines.
--
You ar
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #31 from Jon Ciesla ---
Git done (by process-git-requests).
--
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug
https://bugzilla.redhat.com/token.cgi?t=uPSXKcJ7vr&a=
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Susi Lehtola changed:
What|Removed |Added
Flags||fedora-cvs?
--- Comment #
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Dominik 'Rathann' Mierzejewski changed:
What|Removed |Added
Flags|fedora-review? |
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #28 from Susi Lehtola ---
(In reply to comment #23)
> Actually, I think it's wrong to resolve the file conflicts by changing the
> library SONAMEs. All versions, irrespective of the thread options, implement
> the same
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #27 from Susi Lehtola ---
So, the reasoning boils down to a very simple fact: whether a program will be
SMP is determined at compile time. Thus the choice of the flavor of library to
use should also be done at compile
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #26 from Susi Lehtola ---
Coming from a high-performance computing background, I just listed the reasons
above in comment #24.
Even the environment modules solution is really a pain in the ass, since what
the environm
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #25 from Dominik 'Rathann' Mierzejewski ---
Go with environment modules, then. By using different SONAMEs, you're forcing
users to relink if they want to switch from serial to parallel. Why do you
think compile-time ch
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #24 from Susi Lehtola ---
Then ATLAS does it wrong.
Putting the choice on system level is giving you the choice between a hammer
and a sledgehammer, when you might want to use a hammer for some things and the
sledgeha
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #23 from Dominik 'Rathann' Mierzejewski ---
Created attachment 677085
--> https://bugzilla.redhat.com/attachment.cgi?id=677085&action=edit
parallel installable libopenblas versions with same SONAME
Actually, I think
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #22 from Susi Lehtola ---
(In reply to comment #21)
> Would it be possible to produce a 64bit integer variant say openblas64 by
> adding INTERFACE64=1?
Yes, it would. But I figure the 4-byte integer variant is still
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Robert Szalai changed:
What|Removed |Added
CC||r.sza...@bris.ac.uk
---
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #20 from Susi Lehtola ---
Turns out the problem was that in the OpenMP version you still need
USE_THREADS=1. Fixed in the above.
--
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #19 from Susi Lehtola ---
I've fixed everything else except the unused-direct-shlib-dependency stuff.
I'll contact upstream about it.
http://theory.physics.helsinki.fi/~jzlehtol/rpms/openblas.spec
http://theory.physi
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #18 from Dominik 'Rathann' Mierzejewski ---
Thanks, it builds on EPEL-6 fine now.
rpmlint output for rawhide build:
openblas-threads.x86_64: W: spelling-error Summary(en_US) pthreads -> threads,
p threads, threader
op
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #17 from Susi Lehtola ---
Well, I remembered that lapacke support can be turned off. Here's a version
that builds also in EPEL.
http://theory.physics.helsinki.fi/~jzlehtol/rpms/openblas.spec
http://theory.physics.hels
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #16 from Susi Lehtola ---
Looks like lapacke was integrated in lapack only in version 3.4.0, and it is
thus unavailable in EPEL, which only has 3.2.1 (EL6) vs 3.0 (EL5).
Now this does raise once again the question of
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #15 from Susi Lehtola ---
Nope, no typo - liblapacke contains the C interfaces.
$ rpm -qf /usr/lib64/liblapacke.a
lapack-static-3.4.1-2.fc18.x86_64
--
You are receiving this mail because:
You are on the CC list for
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #14 from Dominik 'Rathann' Mierzejewski ---
Build fails on EL-6/x86_64 at %prep stage:
Executing(%prep): /bin/sh -e /home/rathann/build/tmp/rpm-tmp.MbqSqa
...
+ mkdir netliblapack
+ cd netliblapack
+ ar x /usr/lib64/l
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #13 from Susi Lehtola ---
... but don't include the reference implementations of functions that have an
optimized implementation in openblas.
http://theory.physics.helsinki.fi/~jzlehtol/rpms/openblas.spec
http://theo
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #12 from Susi Lehtola ---
Use system version of LAPACK.
http://theory.physics.helsinki.fi/~jzlehtol/rpms/openblas.spec
http://theory.physics.helsinki.fi/~jzlehtol/rpms/openblas-0.2.5-3.fc18.src.rpm
--
You are receiv
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #11 from Susi Lehtola ---
Dominik - could you approve the review?
--
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug
https://bugzilla.redhat.com/token.cgi?t=CGS52QS
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #10 from Susi Lehtola ---
Oh, I fixed the build issue on i386 and RHEL5 as well.
http://theory.physics.helsinki.fi/~jzlehtol/rpms/openblas.spec
http://theory.physics.helsinki.fi/~jzlehtol/rpms/openblas-0.2.5-2.fc18.sr
Product: Fedora
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #9 from Jussi Lehtola ---
(In reply to comment #8)
> I'll be taking a two-week vacation now, so I won't be able to look at this
> until I get back.
Whoops...
Anyway, I have rebased the package to 0.2.5, otherwise the
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Alex Lancaster changed:
What|Removed |Added
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #8 from Jussi Lehtola 2011-10-13 15:51:48
EDT ---
(In reply to comment #7)
> > > Minimum requirements for Fedora ar
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #7 from Dominik 'Rathann' Mierzejewski
2011-10-13 15:11:04 EDT ---
(In reply to comment #6)
> (In reply to comment
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #6 from Jussi Lehtola 2011-10-05 18:38:54
EDT ---
(In reply to comment #5)
> The gcc command which the library is l
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #5 from Dominik 'Rathann' Mierzejewski
2011-10-05 17:29:01 EDT ---
Builds in mock rawhide/x86_64.
$ rpmlint /var/l
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #4 from Jussi Lehtola 2011-09-21 07:23:44
EDT ---
FYI: I've already asked upstream to soname the libraries properly
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Dominik 'Rathann' Mierzejewski changed:
What|Removed |Added
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #2 from Jussi Lehtola 2011-09-18 18:13:13
EDT ---
I ran a series of matrix diagonalization benchmarks, and on my In
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
Jussi Lehtola changed:
What|Removed |Added
-
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=739398
--- Comment #1 from Jussi Lehtola 2011-09-18 12:37:23
EDT ---
There is one slight problem with the package: -devel picks up a d
58 matches
Mail list logo