Hello Kiel,

This is on purpose (unless the test is explicitly testing multi-core).
The goal is to minimize the variance of micro-benchmarks due to the
effect of turbo-boost, downclocking or even memory bandwidth
congestion. I recall that some tests were very particularly affected
by this. An improved approach could be to study and find such tests
and mark them to run with a "global-lock".

François

On Thu, Nov 19, 2020 at 4:54 PM Kiel Friedt <[email protected]> wrote:
>
> Hi All,
>
>
>
> I am benchmarking using archery and during the initial build I do see many 
> cores are used for building. Averaging around 60% on a 8 core machine. But 
> after building only a single core is used. How can I allow tests to run and 
> utilize all cores on my system?
>
>
>
> Kiel
>
>
>
> IMPORTANT NOTICE: The contents of this email and any attachments are 
> confidential and may also be privileged. If you are not the intended 
> recipient, please notify the sender immediately and do not disclose the 
> contents to any other person, use it for any purpose, or store or copy the 
> information in any medium. Thank you.

Reply via email to