What you said is right, stanford releasing SMP is just for 4cores.
But some said, can execute 2*SMP on 4cores without losses any performance ... :shrug: Yeap, I don't know, this is what I read :D
Printable View
JC to see if the SMP client has issues with so many cores you could try and run in just a Bloomfield CPU and compare the PPD, then we can see if it is a software issue on the F@H client.
I hope you will be able to release more info/pictures soon
Damn, enough with this SMP crap ... Maybe I'll just shift this machine to BOINC dedication ... :shakes:
Don't quite understand, why F@H don't make one like BOINC Client ? ...
http://i245.photobucket.com/albums/g...neous/hung.jpg
...
I'm not exactly an expert but I thought BOINC just distributed different tasks to different cores. SMP tries to leverage parallelism by having four cores work on the same task. You could run 16 or 8 different single-threaded clients with different machine IDs to see what happens then.
What a weird SMP, today 3 & 4 SMP became normal again, continued the previous WU :shrug:
http://i245.photobucket.com/albums/g...ous/FAH_G5.jpg
...
interesting, also i see your running a 2665 WU. On the offical forums the say 2665 WUs cost quite some performance/PPDs.
but this score is still awesome, beats a Q9450 + 8800gts 512 :yepp:
Another result for Chess mania users (TogaII) :shrug: ,
Thanks to someone for this brief analysis :yepp:
http://i245.photobucket.com/albums/g...lem/TogaII.jpg
...
Register at wcgrid.org, join team XtremeSystems, set 16 cores to the default profile (under Profile manager) download BOINC client, and start it. Let's see some 16core goodness :up:
Uh, a single Bloomfield at 2.93 beats two Harpertowns and almost reaches up to Skulltrail at 4.0?
I need a new computer hobby.. Something involving a lot of threads. :cool:
well its dp gainstown so its different form bloomfield anyways. :p:
This is as far as I can give ... :D Gainestown platform ...
http://i245.photobucket.com/albums/g...alem/Neh_6.jpg
...