Date: Tue, 13 Feb 2007 10:54:10 -0800 From: John Bowman <jpbowman@therandomco.com> To: freebsd-ports@freebsd.org Subject: math/ATLAS upgrade fails Benchmarking xzllttst Message-ID: <45D20952.6030503@therandomco.com>
next in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------030405030108010401040700
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
I am duplicating Karsten Rothemund's earlier post, as my error message
is identical except for timing numbers:
Unfortunately I only have the last few lines (because of the portupgrade):
10 cases: 10 passed, 0 skipped, 0 failed
Benchmarking xcllttstF
NREPS UPLO N lda TIME MFLOPS RESID
===== ===== ===== ===== ============ ============ ============
1 Lower 100 100 0.00000 0.000 6.688674e-03
1 Lower 200 200 0.02719 396.642 4.398721e-03
1 Lower 300 300 0.07179 505.188 2.341770e-03
1 Lower 400 400 0.15059 569.836 2.147220e-03
1 Lower 500 500 0.28906 579.186 2.432632e-03
1 Lower 600 600 0.48022 601.979 1.839629e-03
1 Lower 700 700 0.76107 602.840 1.436337e-03
1 Lower 800 800 1.13554 602.873 1.535296e-03
1 Lower 900 900 1.59588 610.590 9.815540e-04
1 Lower 1000 1000 2.18219 612.381 1.306027e-03
10 cases: 10 passed, 0 skipped, 0 failed
Benchmarking xzllttst
NREPS UPLO N lda TIME MFLOPS RESID
===== ===== ===== ===== ============ ============ ============
assertion ATL_zpotrf(CblasColMajor, Uplo, N, A, lda) == 0 failed, line 344 of file ../llttst.c
*** Error code 255
Stop in /usr/ports/math/atlas.
*** Error code 1
Stop in /usr/ports/math/atlas.
** Command failed [exit code 1]: /usr/bin/script -qa /tmp/portupgrade.72539.60 e
nv UPGRADE_TOOL=portupgrade UPGRADE_PORT=atlas-3.6.0,1 UPGRADE_PORT_VER=3.6.0,1
make
** Fix the problem and try again.
It was in the THREADED_PIC run of the compilation.
Since this appears to be an error in the benchmark / test software
rather than in ATLAS proper, I'm thinking of removing the test from the
Makefile and trying again. However, doing this sort of thing is more an
act of desperation than a well-thought-out strategy for success.
Any suggestions would be greatly appreciated.
John
--------------030405030108010401040700--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?45D20952.6030503>