Wow. GPU to 72 is the #2 team overall, and #1 for TF and P-1!:whistle::grin:
[url]http://www.mersenne.org/report_top_teams/[/url]
[url]http://www.mersenne.org/report_top_teams_TF/[/url]
[url]http://www.mersenne.org/report_top_teams_P-1/[/url]

[QUOTE=kladner;305303]Wow. GPU to 72 is the #2 team overall, and #1 for TF and P-1!:whistle::grin:
[url]http://www.mersenne.org/report_top_teams/[/url]
[url]http://www.mersenne.org/report_top_teams_TF/[/url]
[url]http://www.mersenne.org/report_top_teams_P-1/[/url][/QUOTE]

Yeah, but neither of the other two top-3 teams are major TFers, they all use only CPUs...

[QUOTE=Dubslow;305305]Yeah, but neither of the other two top-3 teams are major TFers, they all use only CPUs...[/QUOTE]

True. But P-1 is CPU work. The team seems to be pretty good on both fronts.

[QUOTE=LaurV;304280]WHAT? When this happened? :rant:
I think you better have #4 in your sight, if you can't get #2 this week. I didn't do DC for a while, and now I am still dealing with "332M to 70 bits", but in no more then 5-6 days I would target to shot chalsall in the leg by myself, hehe...
So don't be so happy :razz:[/QUOTE]

Hmm... I get 8-9 ms/iter on one core, but when I put all four cores to task they all shoot up to 13 ms/iter (and my scrolling gets a wee bit laggy)... derned memory bottlenecks. Let's see what I can do though. :smile:

[QUOTE=Dubslow;306269]Hmm... I get 8-9 ms/iter on one core, but when I put all four cores to task they all shoot up to 13 ms/iter (and my scrolling gets a wee bit laggy)... derned memory bottlenecks. Let's see what I can do though. :smile:[/QUOTE]
Have you tried testing 2 core/exponent to see if 2 doubles are faster than 4 singles?

[QUOTE=bcp19;306289]Have you tried testing 2 core/exponent to see if 2 doubles are faster than 4 singles?[/QUOTE]
I seem to recall having tried that, without much success. I'll try again though in a day or two.

[QUOTE=Dubslow;306309]I seem to recall having tried that, without much success. I'll try again though in a day or two.[/QUOTE]
I only mention it cause, IIRC, Core 0/1 share an L2 cache as do Core 2/3. I know on the machine I am doing the 332M exp that when I was running Core 0 on it, Core 1 and 2 on DC/LL and 3 on mfaktc (GTX 560). The 332M was ~180ms in this setup. I then got a 480, so put a second core on mfaktc and noticed it was now hitting ~168ms/iter. So, I played around some more and I ended up stopping Core 1 and the time dropped to ~159ms. Curious I set it to 0/1 on the 332M and it's currently running 75.5ms.

AFAIK this is only on core 2.

[QUOTE=petrw1;304323]Of course we are all assuming I don't have a secret weapon just waiting to be unleashed.... [/QUOTE]
Grrrr... you had one, eventually... How many cores were there reporting in the same time, for that jump? I have the right to know what I am up against :razz:
Or maybe I had a too late start... But I am still in the game... during the weekend I moved 4 cores of i7-2600k and 4 cores of xeon from LL and P-1 to DC (the effect of this movement will not be seen immediately, but it will be visible in few days).

[QUOTE=LaurV;306413]Grrrr... you had one, eventually... How many cores were there reporting in the same time, for that jump? I have the right to know what I am up against :razz:
Or maybe I had a too late start... But I am still in the game... during the weekend I moved 4 cores of i7-2600k and 4 cores of xeon from LL and P-1 to DC (the effect of this movement will not be seen immediately, but it will be visible in few days).[/QUOTE]

Well actually I have 16 cores on GPU72 DC all the way from a PIV to i5.
Currently, my throughput is 2 or 3 a day.

:sparta::ouch::farley::fight:
:offensive::hamster::tank:
:choppa::kitten:

[QUOTE=chalsall;283095]TF work expires after a month unless an extension is requested. This is clearly stated on the assignments page.[/QUOTE]
Is this number still valid? I don't remember what it said on the assignments page, and when I try to go there, I get [QUOTE="http://www.gpu72.com/account/getassignments/lltf/"]Sorry Patrik Johansson, but you already have too many assignments.

Since you joined the project you have done on average 124.75 GHz Days of Trial Factoring work per day.

You currently have 133 assignments totalling 3237.70 GHz Days of work assigned, or 25 days worth based on your history. The oldest is 19 days old.[/QUOTE]