[Numpy-discussion] numpy blas running slow: how to check that it is properly linked

David Cournapeau cournape@gmail....
Tue Sep 20 12:18:25 CDT 2011


On Tue, Sep 20, 2011 at 9:56 AM, David Cottrell
<david.cottrell@gmail.com> wrote:
> Thanks, just getting back to this. I just checked again, and after
> setting my LD_LIBRARY_PATH properly, ldd shows _dotblas.so pointing
> and the sunmath and sunperf libraries. However the test_03.py still
> runs at about 8-9 seconds ... far too slow.
>
> ~/local/lib/python3.1/site-packages/numpy/core $ ldd _dotblas.so | sed
> -e 's/$me/$USERNAME/g'
>        libsunperf.so.8 =>
> /home/$USERNAME/local/archive/SolarisStudio12.2-solaris-sparc-tar-ML/solstudio12.2/lib//libsunperf.so.8
>        libsunmath.so.1 =>
> /home/$USERNAME/local/archive/SolarisStudio12.2-solaris-sparc-tar-ML/solstudio12.2/lib//libsunmath.so.1
>        libgcc_s.so.1 =>         /usr/sfw/lib/libgcc_s.so.1
>        libfsu.so.1 =>
> /home/$USERNAME/local/archive/SolarisStudio12.2-solaris-sparc-tar-ML/solstudio12.2/lib//libfsu.so.1
>        libfui.so.2 =>
> /home/$USERNAME/local/archive/SolarisStudio12.2-solaris-sparc-tar-ML/solstudio12.2/lib//libfui.so.2
>        libpicl.so.1 =>  /usr/lib/libpicl.so.1
>        libmtsk.so.1 =>  /lib/libmtsk.so.1
>        libm.so.2 =>     /lib/libm.so.2
>        libc.so.1 =>     /lib/libc.so.1
>        libm.so.1 =>     /lib/libm.so.1
>        libdl.so.1 =>    /lib/libdl.so.1
>        libdoor.so.1 =>  /lib/libdoor.so.1
>        libthread.so.1 =>        /lib/libthread.so.1
>        libkstat.so.1 =>         /lib/libkstat.so.1
>        libpthread.so.1 =>       /lib/libpthread.so.1
>        librt.so.1 =>    /lib/librt.so.1
>        libaio.so.1 =>   /lib/libaio.so.1
>        libmd.so.1 =>    /lib/libmd.so.1
>        /platform/SUNW,Sun-Fire-V490/lib/libc_psr.so.1
>        /platform/SUNW,Sun-Fire-V490/lib/libmd_psr.so.1
>
> ~/local/lib/python3.1/site-packages/numpy/core $ ~/python/numpy/B/test_03.py
> No ATLAS:
> 8.49377894402
> (1000, 1000) (1000,) (1000, 1000)

As mentioned earlier, we currently only optimize dot if you are using
ATLAS, not any generic BLAS implementation. Your linear algebra
functions will be fast, though,

cheers,

David


More information about the NumPy-Discussion mailing list