Hi,
Sean has left the following comment at Investigate performance of setting
thread affinity
http://www.google-melange.com/gci/task/view/google/gci2012/7960222:
hot diggity
So Skriptkid, a little encouragement perhaps. I applied your patch and did
a real quick-fix for the various build issues I alluded to (which you
should still fix properly). I then tested the performance impact on a real
compute system.
On a 64-processor POWER7 beast, I'm seeing our *debug* benchmark results
going from around 44000 VGRs to around 67000 VGRs. That's more than a 50%
performance increase, but wait, it gets better...
Our *optimized* production benchmark results increase from around 121000
VGRs to 185000 VGRs!
That. Is. Awesome.
Your patch is going to make some big attention in our next major release.
I'll also be adding extra tasks to implement thread affinity for BSD and
Windows.
Greetings,
The Google Open Source Programs Team
---
You are receiving this message because you are subscribed to Investigate
performance of setting thread affinity.
To stop receiving these messages, go to:
http://www.google-melange.com/gci/task/view/google/gci2012/7960222.
------------------------------------------------------------------------------
LogMeIn Rescue: Anywhere, Anytime Remote support for IT. Free Trial
Remotely access PCs and mobile devices and provide instant support
Improve your efficiency, and focus on delivering more value-add services
Discover what IT Professionals Know. Rescue delivers
http://p.sf.net/sfu/logmein_12329d2d
_______________________________________________
BRL-CAD Tracker mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-tracker
