Looking to benchmark pytorch models on Core processors

[mike02]

Hello

I am looking to benchmark our Pytorch models across the Core series of processors to understand the cost / benefit tradeoff. 

I only see a i9-11900 Processor in the list of nodes (generated via pbsnodes command) on devcloud. There are also a few Xeon processors but these have a very poor cost / benefit tradeoff for our models. 

Do the list of processors change over time or is there a way to run benchmarks on generation 12 / 13 - i7 and i9 processors?

Thanks 

Michael

(By the way -  this is a great platform for being able to quickly check performance)

[AthiraM_Intel]

Hi,

Thank you for posting in Intel Communities.

For your information, current Terms and Conditions of using DevCloud do not allow benchmarking or publishing any related results.

For more information on the Terms & Conditions of DevCloud usage refer https://www.intel.com/content/www/us/en/legal/terms-of-use.html , Section 9.2 prohibits publishing performance benchmarks.

You could check the available platforms in DevCloud for oneAPI using the below command :

pbsnodes | grep properties | sort | uniq -c

If this resolves your issue, make sure to accept this as a solution. This would help others with similar issue. 

Thank you

[mike02]

Thanks for the reply and information. The central part of my question was if the list of processors available on the platform rotates over time. My takeaway from the reply is that the list does not change over time.

Thanks

[AthiraM_Intel · Accepted solution]

Hi,

We are sorry for the confusion, processors available on DevCloud for oneAPI do change over time. We will add or deprecate some processors over time.

If you have any further issue, please let us know.

Thanks

[AthiraM_Intel]

Hi,

Thanks for accepting our solution. If you need any additional information, please post a new question as this thread will no longer be monitored by Intel.

Thanks