On Wed, Oct 7, 2015 at 7:44 PM, Petr Baudis <[email protected]> wrote: > On Wed, Oct 07, 2015 at 02:29:27PM +0200, Erik van der Werf wrote: >> A measure that I find reasonable is a limit on number of threads x >> clock frequency.
> I'm not sure this would work well. The #playouts difference between > an old Bulldozer and new Haswell with the same nominal frequency and > #cores might be many tens of percents. Another question is how to find > the scaling factors for other architectures? Instead of frequency, one could use the single threaded Fhourstones score. https://en.wikipedia.org/wiki/Fhourstones http://tromp.github.io/c4/fhour.html regards, -John _______________________________________________ Computer-go mailing list [email protected] http://computer-go.org/mailman/listinfo/computer-go
