Re: [Boost-testing] Sandia intel & pathscale

From: K. Noel Belcourt (kbelco_at_[hidden])
Date: 2007-12-10 13:08:34

Hi Rene,

On Dec 10, 2007, at 9:23 AM, Rene Rivera wrote:

> It seems there are some problems on those two machines. I think the
> config problem with intel was mentioned:
>
>    Compiler driver / backend version mismatch: unset != 900
>
> But we also now have some backend issues:
>
>    icpc: error: could not find directory in which g++ resides
>
> And on the pathscale machine some disk space problems:
>
>    /var/scratch2/boost/boost/libs/config/test/abi/abi_test.cpp:0:  
> fatal
> error: when writing output to /tmp/ccii.lUgIG3: No space left on  
> device

We had an network and disk failure outage since last Friday.  Things  
are slowly returning to normal (i.e. it may take another day before  
I'm able to run tests, etc...).

The Intel 9.0 problem is, I think, an issue with running multiple  
toolsets at the same time.  For example, my run script is

run.py ... --toolsets="intel-8.1 intel-9.0 intel-9.1"

and my LD_LIBRARY_PATH is setup for intel-9.1.  So I'm guessing that  
I need to run intel-9.0 in a separate run.py so I can set the  
LD_LIBRARY_PATH correctly for intel-9.0.

-- Noel