or Create a profile
25 Feb, 2016 10:11 AM
Most of our workload is from single-core jobs running on a cluster. So for us, an appropriate benchmark ("Geekbench parallel mode") would be:
- driver checks machine specs, finds N cores
- driver starts N copies of the single-threaded Geekbench (ie does not run the multithreaded bits)
- for best results, the various pieces of each of the N would be run in random order, that is different order for each copy
the randomization is meant to emulate single-user jobs with uncorrelated workloads.
Is there a mode like this or could you make it?
Formatting help /
(switch to plain text)
(switch to Markdown)
You can attach files up to 10MB
If you don't have an account yet, we need to confirm you're human and not a machine trying to post spam.
A conversation has been started with the Primate Labs staff to resolve this discussion.
This discussion is private.
Only you and Primate Labs support staff can see and reply to it.
This discussion is public. Everyone can see and reply to it.
You can use Command ⌘ instead of Control ^ on Mac
Powered by Tender™.