Message boards :
Number crunching :
Observation of CreditNew Impact (2)
Message board moderation
Previous · 1 . . . 17 · 18 · 19 · 20
Author | Message |
---|---|
Eric Korpela Send message Joined: 3 Apr 99 Posts: 1382 Credit: 54,506,847 RAC: 60 |
Eric, is it possible that the differences in credit rates for MB and AP cpu are due to different hardware populations? IE people with higher-end hardware choose not to run AP? It's possible. The differences between CUDA5 and CUDA2.3 are significantly due to hardware differences. I haven't done statistics on how many people have turned off either astropulse or setiathome. If I get time, maybe I'll try. @SETIEric@qoto.org (Mastodon) |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
I am assuming 1 per GPU. I assume that accounts for the majority of our users and result. That could be one of the sources of the diferences, most of the big crunchers runs 2 or more WU MB and normaly only 1 AP at a time, at least those who have a NV fermi/keppler based or newer ATI GPUs, as most of the 100 top crunchers have just look the top hosts lists. Sure they have a very large contribution on the GPU´s crunched totals. |
Thomas Send message Joined: 9 Dec 11 Posts: 1499 Credit: 1,345,576 RAC: 0 |
Thanks very much Eric for taking the time and trouble to respond with as much precision. Really appreciated. It's always good to have answers. So we have to get used to this new situation. Sniff... |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
In Astropulse the GPU routines were written directly in OpenCL. Part was written in Brook+ and ported to OpenCL too. EDIT: and indeed that table includes only credits per app. W/o hardware classes. For example, ati_cal gives very low credits per hour. What it means? It just means that app mostly run on very old and slow hardware (as should be). SETI apps news We're not gonna fight them. We're gonna transcend them. |
Blurf Send message Joined: 2 Sep 06 Posts: 8962 Credit: 12,678,685 RAC: 0 |
Every credit method we've ever used, with the exception of FLOP counting punished optimization, especially optimization of the stock apps. You're not an employee--you're not paid-you choose to run Seti. Please stop using the analogy. |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Every credit method we've ever used, with the exception of FLOP counting punished optimization, especially optimization of the stock apps. Sure you don´t read the entire thread, Eric explain why all things happening and now we move to a next level, try to find way the diference apears and if possible find a fix for that. |
Ron Send message Joined: 24 Aug 99 Posts: 42 Credit: 34,544,679 RAC: 0 |
This is another way this glitch is affecting people: This computer never runs out of AP work and has no concerns about the V7 credits. His next step is, I expect, to hide his computers so no one will know what he's doing. All tasks for computer 6016862 State: All (752) · In progress (391) · Validation pending (261) · Validation inconclusive (13) · Valid (86) · Invalid (0) · Error (1) Application: All (752) · AstroPulse v6 (703) · SETI@home Enhanced (5) · SETI@home v7 (44) |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
This is another way this glitch is affecting people: Read this it´s all about his owner, hope he allready stops to recheduling. http://setiathome.berkeley.edu/forum_thread.php?id=72502 Yes that another good reason why someone must fix that soon. But give some time to Eric, he post why the diference is happenings and belive he´s allready working on that. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.