| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Human Proteome Folding - Phase 2
Thread: Shortchanging on points for HPF Phase2 on Mac? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 3
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi,
This may well be not the right forum, but here goes. Basically, my brand new 24" iMac seems to claim way more points than it is granted, but only for the HPF work units. For the other projects it normally underclaims and gets awarded above the amount expected. Consequently, I'm feeling rather short-changed for HPF work units. For instance, on a recent HPF work unit I claimed/granted: 154.1 / 111.8 pts (7.28 hours CPU) and another example: 147.0 / 115.6 pts (7.02 hours CPU) This isn't happening for my other projects... e.g. for a recent HCC work unit, I claimed/granted 98.7/110.9 for 4.71 hours of CPU time. So... it seems to me that HPF is giving a very unfair allocation of points for my Mac, and I should be granted nearer to what I'm claiming in order to bring things in line with what I'm earning per CPU hour on other projects. Note that I do not observe this disparity on the work units returned by my PC. I know points aren't really that important in the grand scheme of things, but any idea why this is so? Cheers, Rob |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Sorry, you were saying?
----------------------------------------lp986_ 00003_ 2-- 361642 Valid 03/21/2008 12:27:51 03/26/2008 21:15:07 9.28 151.2 / 143.1 lp973_ 00029_ 5-- 361642 Valid 03/21/2008 03:05:49 03/26/2008 09:10:28 8.29 135.0 / 129.1 lp971_ 00009_ 0-- 361642 Valid 03/21/2008 02:12:27 03/25/2008 23:59:25 8.23 134.1 / 116.9 lp965_ 00011_ 5-- 361642 Valid 03/21/2008 01:52:59 03/25/2008 21:45:35 7.90 128.6 / 124.7 lp970_ 00004_ 9-- 361642 Valid 03/21/2008 01:21:41 03/25/2008 21:36:26 8.24 134.2 / 125.1 lp971_ 00008_ 4-- 361642 Valid 03/21/2008 00:48:12 03/25/2008 19:39:55 7.70 125.4 / 120.8 lp965_ 00010_ 10-- 361642 Valid 03/20/2008 23:05:43 03/25/2008 15:14:28 8.07 131.4 / 125.1 lp954_ 00015_ 5-- 361642 Valid 03/20/2008 17:03:50 03/25/2008 13:16:21 8.39 136.6 / 119.7 lp938_ 00028_ 13-- 361642 Valid 03/20/2008 07:54:13 03/24/2008 18:17:37 6.92 112.7 / 107.7 lp897_ 00031_ 13-- 361642 Valid 03/19/2008 09:03:53 03/24/2008 16:36:48 8.46 137.8 / 128.0 lp900_ 00000_ 14-- 361642 Valid 03/19/2008 09:01:21 03/24/2008 10:02:53 6.97 113.5 / 110.4 lp899_ 00036_ 0-- 361642 Valid 03/19/2008 09:01:21 03/24/2008 09:05:02 8.05 131.0 / 129.1 lp883_ 00027_ 13-- 361642 Valid 03/18/2008 23:48:06 03/24/2008 09:02:00 7.60 123.7 / 118.0 lp859_ 00046_ 7-- 361642 Valid 03/18/2008 08:11:58 03/23/2008 09:02:57 7.59 124.1 / 118.6 lp848_ 00014_ 10-- 361642 Valid 03/17/2008 23:51:44 03/22/2008 22:52:26 8.98 146.7 / 139.0 Seems the going rate is 110-120 on average. The claims per hour computed are absolutely identical and you can check for yourself if not done already. It's just that from our experiences with other projects it is a known fact that one computer/OS combo is not as efficient at doing a task as the other, particular if there is a fare bit of page faulting involved. We hear few about getting 10-20 more over the claim, but I see it frequently, so on the balance, if you feel your particular machine is short changed too much, take the project out. That said, WCG is preparing the science apps for readiness to run with BOINC 6. Some was learned from the excessive page faulting on HCC jobs, but again only for some OS/Hardware combo's, so it's not unlikely that knowledge is applied in future releases. After all, optimized sciences run faster and allow for projects to complete quicker. On the closing, i know you have a 64bit OS and probably a 64bit client. They have a way overrated Integer benchmark (Dhrystone) compared to the 32bit science execution, thus think in fact where larger groups meet up to make up a quorum 15 average you get more of the 32bit credit claim ingredient
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for the info. I notice I do seem to be getting a lot of page faulting. And some other weird stuff going on too. But I guess that's a thread for another sub-forum.
|
||
|
|
|