A Monday 23 May 2011 10:38:29 Rolv Seehuus escrigué:
> Hi all,
> 
> First things first: this is my first posting to this list. I use
> PyTables for building hdf5 datasets, for visualization purposes.
> 
> I run in to a build-issue when installing PyTables 2.2.1, and wonder
> if there is a known resolution for this (nothing in the faq)
> 
> I need to be able to build and deploy my application on CentOS 4.8,
> so we have previously used quite old versions of the needed
> toolchain. I want to upgrade it, so I have installed Python2.6,
> Numpy 1.5.1, Cython 0.14.1 and numepr 1.4.2. All works good so far.
> 
> When building PyTables there is a build error in blosc.c. It
> complains that it cannot find pthread_barrier_t and friends. I was
> able to make everything build by commenting out the define for
> _POSIX_BARRIERS_MINE - from the looks of it, it seems that the code
> then falls back to using mutexes and condvars instead. Running the
> tests, and all looks okay. It seems that the pthreads headers on
> CentOS4.8 is misinforming about it's capabilities.
> 
> So then to my question(s):
> - Is this a known issue on this configuration? (It is obscure enough
> to not warrant a FAQ I suppose.)

No, this is unknown.  But probably you are right: your CentOS box seems 
to be lying when it says that it supports _POSIX_BARRIERS with a value 
larger than 20012L.  Maybe you should file a bug for CentOS.

> - And is this safe? (As in, are there other configurations that will
> build with mutexes/condvars that are in widespread use)

Yes, your solution is completely safe, as it fall backs to simulate 
barriers using mutex locks and conditions.  My benchmarks also says that 
the difference in performance among the two methods is generally 
negligible.

-- 
Francesc Alted

------------------------------------------------------------------------------
Simplify data backup and recovery for your virtual environment with vRanger. 
Installation's a snap, and flexible recovery options mean your data is safe,
secure and there when you need it. Data protection magic?
Nope - It's vRanger. Get your free trial download today. 
http://p.sf.net/sfu/quest-sfdev2dev
_______________________________________________
Pytables-users mailing list
Pytables-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/pytables-users

Reply via email to