Pretty informative Video actually... Thank you for posting! 

Linux GPU core?....Native linux folding?...no wine?
120k on titan
110k on 7970
81K on 680
 
to come:
Intel Xeon Phi compatiblility
Intel GPU compatibility
more speed = More points
 
so where is the new core?
Id like to see it
 
Not many internal testers from the EVGA team...do we have any?
 
Questions?
Reddit IAMA
www.reddit.com/r/folding
 
 
OOOHHHH!!!! core 17 is back on beta

proteneer
We released OpenMM 5.1 based core 17 today
AMD Linux - not supported yet - linux driver unstable
....
 With the native linux core, you won't need to use wine anymore.
 

Linux GPU cores start today
 
Thx you proteneer for keeping us more informed...this is what we like to see more from stanford
if you need help with the core or openMM(coding) - I might be able to help in the coming months
 

P7663 Beta testing for FAHCore_17 and OpenMM5.1  

proteneer

I will try to improve my Engrish in the future, in the mean time, there's Closed Captions on the Youtube Video =)

It could be my hearing more than your speaking!
 
Cannot thank you enough for keeping us up to date. The direct result is it makes "Many" of us pump more into F@H. Thank you! 

if you want to apply for Beta Team
http://foldingforum.org/viewtopic.php?f=16&t=8
 
 
read the requirements and then post in the thread your system specs 
 
It be helpful to acknowledge you read and agree to the conditions
 
 
 
Update on estimated PPD on new Beta WU
670FTW 70K on 7663
570 31K on 7663
 
WOW!!.... if this holds My PPD just doubled
seems the 570 doesn't get much help with the new core
 

FAHBench and Freenode

Be warned!!!
core 17 still uses a lot of CPU cycles with Nvidia cards...so keep a few threads avail for the GPUs

It be the same as last core 17... just leave 2 cores/threads for all your GPUs 
 
1 GPU is about 5%-10% CPU (3930k)
I'll let you know more as I get them loaded up with the new WU's
 
I have 1 old WU's being serviced still
 
 
my core2 quad ..it's almost using a full core for 1 GPU
 
my 3930k is now up to 15-20% CPU usage with two GPU's

On my SR-2 (with dual, 4-physical-core Xeons), I had to leave a minimum of (4) threads free to avoid having crashes while Folding on my GTX690.  I had tried (2) threads and was still getting crashes that were killing the GPU Folding WU too.  I'll have to give this latest core and 7663 WU a try just to see if anything has improved in that regard. 
 
I tell ya fellas, this is the kind of interaction and sharing of information that we have wanted to get from Stanford for a long time.  I feel that we have seen a turn for the better in that regard with the combination of Blog posts, Proteneer's direct involvement in Forums and now this video which was very interesting to watch.  Looking good PG...well done, thank-you! 

is this a hint?

zdanger
So we're talking 110k PPD for a 7970 at ~270 watts? PPD/watt isn't too bad... Still not as good as my bigadv rigs.

proteneer
 
Wait until -biggpuadv :P

 

proteneer
after we do some more testing on 500k+ atoms.

Recent post on the AMA:
 

proteneerOSX won't be supported because their OpenCL implementation is kinda borked. Linux is about 5-7% faster than Windows because the drivers don't need to deal with the Windows way of doing things.

slapdashbr
 
My question: do you think there is anything the reddit community can do to promote folding@home and other distributed computing projects?


permalink

proteneer F@H GPU Core Programmer[S] 
 
It never hurts to tell more people about [link=mailto:Folding@home]Folding@home[/link]. Get more involved in internal testing in our IRC channel. Also, you guys can tell NVIDIA to fix their OpenCL drivers, and AMD for improve their OpenCL linux drivers. In terms of concrete things, organizing more events such as competitions, etc. is always good.

Afterburner

FYI... Mine is using "Three" cores... (Bottom right in the pic showing 33% CPU Usage)

920 DO with (2) 680's in SLI...

mines at 25-28% with three 670FTW - that be about 3 threads... 1 thread per GPU

the new core17 works with HFM again!
7970's @1025 are getting 45-46kppd ea, which is up about 7k from 7662!

 
im so happy now that they arnt sitting on the shelf rusting away!
 
edit: I just fell out of my chair 680's are getting 75k ea! its gonna be a heat wave!

wrinvert

the new core17 works with HFM again!
7970's @1025 are getting 45-46kppd ea, which is up about 7k from 7662!

Project ID: 7663
Core: ZETA
Credit: 1600
Frames: 100


Name: HD7970 Slot 00
Path: localhost-36330
Number of Frames Observed: 24

Min. Time / Frame : 00:01:16 - 124,561.0 PPD
Avg. Time / Frame : 00:01:18 - 119,801.1 PPD
Cur. Time / Frame : 00:01:16 - 123,692.1 PPD
R3F. Time / Frame : 00:01:18 - 119,349.3 PPD
All Time / Frame : 00:01:18 - 119,349.3 PPD
Eff. Time / Frame : 00:01:20 - 115,256.4 PPD