[Numpy-discussion] Tensor Contraction (HPTT) and Tensor Transposition (TCL)

Paul Springer pavdev at gmx.de
Wed Aug 16 18:29:53 EDT 2017


Am 8/16/17 um 5:38 PM schrieb Stephan Hoyer:
> On Wed, Aug 16, 2017 at 2:39 AM, Paul Springer <pavdev at gmx.de
> <mailto:pavdev at gmx.de>> wrote:
>
>
>>     What version of Numpy are you comparing to? Note that in 1.13 you
>>     can enable some optimization in einsum, and the coming 1.14 makes
>>     that the default and uses CBLAS when possible.
>     I was using 1.10.4; however, I am currently running the benchmark
>     with 1.13.1 and 'optimize=True'; this, however, seems to yield
>     even worse performance (see attached).
>     If you are interested, you can check the performance difference
>     yourself via: ./benchmark/python/bechmark.sh
>
>
> This sounds like you may be using relatively small matrices, where the
> overhead of calculating the optimal strategy dominates. Can you try
> with a few bigger test cases?
>
The sizes of the tensors varies form ~5MB up to ~100MB towards the far
right of the plot; this corresponds to matrices of size ~1000^2 to
~5000^2, thus the sizes should be large enough to amortize any overhead
associated to calculating the optimal strategy.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20170817/1aed124d/attachment.html>


More information about the NumPy-Discussion mailing list