[SciPy-user] Confusing BLAS/LAPACK situation on Mac OS X

Brian Granger bgranger at scu.edu
Sat Jan 28 15:23:42 CST 2006


On Jan 28, 2006, at 1:06 PM, Robert Kern wrote:

Brian Granger wrote:
Thus, It looks like the Apple supplied cblas and clapack libraries are never used, even though they are found at build time.  What is the current status of using Apple's cblas and clapack?  Can I use them from within scipy/numpy?  How?

AFAICT, Accelerate.framework does not include CLAPACK versions of their LAPACK
functions, although they do include CBLAS versions of their BLAS functions. I've
never looked into how scipy determines the existince of the CBLAS routines
because I've always used ATLAS since Accelerate.framework does not contain the
CLAPACK versions.


What about:

/System/Library/Frameworks/Accelerate.framework/Frameworks/vecLib.framework/

Versions/Current/libLAPACK.dylib

and

Headers/clapack.h


Also, Apple's docs on Accelerate claim that they include clapack.

Are these incomplete/unusable for some reason?  It seems silly to build ATLAS if it is already there.







However, it looks like the culprit is line 82 in scipy/Lib/linalg/setup.py.

...
82        if name[0]=='c' and atlas_version is None and newer(__file__,target):
              f = open(target,'w')
              f.write('python module '+name+'\n')
              f.write('usercode void empty_module(void) {}\n')
...

I don't think atlas_version gets set when the ATLAS is Accelerate.framework.


Maybe I will play around with this, but I am not very familiar with how libalg is setup, so I don't know if I will get anywhere.



-- 
Robert Kern
robert.kern at gmail.com

"In the fields of hell where the grass grows high
 Are the graves of dreams allowed to die."
  -- Richard Harter

_______________________________________________
SciPy-user mailing list
SciPy-user at scipy.net
http://www.scipy.net/mailman/listinfo/scipy-user


Brian Granger, Ph.D.
Assistant Professor of Physics
Santa Clara University
bgranger at scu.edu
Phone: 408-551-1891
Fax: 408-554-6965

This message scanned for viruses and SPAM by GWGuardian at SCU (MGW1)



More information about the SciPy-user mailing list