Failure of 't-fat' test in 'make check' in version 5.0.2 ...

[Richard Walsh]

TorBjorn wrote:

>How can a gcc patch be relevant to an issue with icc and gmp?

Did I suggest that ... ??  ... I need to work on my Norwegian ...

What I was suggesting is that this test has been an issue with other compilers
(gcc) and has required patches/fixes (the t-fat test, not the compiler) to make
it work. Looking at such fixes (to the test) might be useful in diagnosing this
issue.

Here is something potentially relevant from the archives:

http://gmplib.org/list-archives/gmp-bugs/2010-January/001813.html

If none of this simulates any interest (I know that bears are start to
hibernate at this time of year), I will jump in an figure it out myself.

Best,

rbw

Richard Walsh
Parallel Applications and Systems Manager
CUNY HPC Center, Staten Island, NY
W: 718-982-3319
M: 612-382-4620

Suppose we were to attribute to nature the property of producing like effects under like circumstances; just such like circumstances we should not know how to find. In truth nature exists once only. It is our schematic mental imitation of nature alone produces like events.   -- Ernst Mach, 1888

________________________________________
From: tg at gmplib.org [tg at gmplib.org]
Sent: Tuesday, November 08, 2011 5:31 AM
To: Richard Walsh
Cc: gmp-bugs at gmplib.org
Subject: Re: Failure of 't-fat' test in 'make check'  in version 5.0.2 ...

Richard Walsh <Richard.Walsh at csi.cuny.edu> writes:

  Not a MAC ... nope this is the head node of a SGI ICE cluster
  with an Intel Nehalem processor.  The compiler is:

SGI Internal Compiler Error cluster, an interesting name.  :-)

  The code makes cleanly.  We build everything in 64-bit by default.
  This is an HPC Center sight.  The actual message from the 'make check'
  output is:

  PASS: t-asmtype
  PASS: t-aors_1
  PASS: t-divrem_1
  /bin/sh: line 5: 20756 Segmentation fault      (core dumped) ${dir}$tst
  FAIL: t-fat
  PASS: t-get_d
  PASS: t-instrument
  PASS: t-iord_u
  PASS: t-mp_bases

  Any further thoughts ... I see that this error has come up before
  with gcc and patches have been provided to yield a fix.  Also, I
  have seen that --disable=static has been suggested.

How can a gcc patch be relevant to an issue with icc and gmp?

--
Torbjörn

________________________________

Change is in the Air - Smoking in Designated Areas Only in effect.<http://www.csi.cuny.edu/tobaccofree>
Tobacco-Free Campus as of July 1, 2012.