Changes between Initial Version and Version 1 of Ticket #17555, comment 8


Ignore:
Timestamp:
May 8, 2025, 5:17:51 PM (6 months ago)
Author:
Tom Goddard

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #17555, comment 8

    initial v1  
    33Details: 4 Intel Xeon Gold 6226R CPU @ 2.90GHz (each 16 cores), watson.cgl.ucsf.edu, 9moj, 1098 seconds = 18.3 minutes, 2600-2800% CPU reported by top, 12.5G VIRT, 6.8G RSS.
    44
    5 Given that it used about 28 cores versus the i7 test that used 12 cores it seems this  Intel Xeon from 2020 is slower than the i7 per core.
     5Given that it used about 28 cores versus the i7 test that used 12 cores it seems this  Intel Xeon from 2020 is slower than the i7 per core.  I did not install the Intel Extensions for PyTorch which are supposed to use AVX-512 CPU instructions, so probably this result just means the PyPi default linux pytorch (which includes cuda support) does not have these Intel-specific optimizations.