Message boards :
Number crunching :
Resource share question
Message board moderation
Author | Message |
---|---|
MattDavis Send message Joined: 11 Nov 99 Posts: 919 Credit: 934,161 RAC: 0 |
It looks like we can get back to asking number crunching questions. I have dabbled in most projects, as my sig shows, but now stick to 5: Seti, Predictor, Einstein, Rosetta, and LHC. All have an equal share on all computers. Now, my question is, since all have an equal share, why is Einstein's RAC so low? LHC's low RAC makes sense, since it is on and off. But shouldn't the RAC of the other 4 be relatively equal too? I give them all equal share so I can do equal science for all, but the differing RAC indicates something might be different. ----- |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
Just use one of Akosf's optimized crunchers for Einstein and your RAC will go up like a skyrocket ;) Aloha, Uli |
MattDavis Send message Joined: 11 Nov 99 Posts: 919 Credit: 934,161 RAC: 0 |
Well, I'm not looking to increase my RAC. Instead, I just use my RAC as an indicator of the proportion of science I'm getting done across projects. ----- |
Saenger Send message Joined: 3 Apr 99 Posts: 2452 Credit: 33,281 RAC: 0 |
Just use one of Akosf's optimized crunchers for Einstein and your RAC will go up like a skyrocket ;) That wasn't the question. 4 projects, all stock app running on stock BOINC should claim all the same credit/h and thus get the same RAC. That's what claimed credit calculation is supposed to do. If it doesn't, something's wrong. Gruesse vom Saenger For questions about Boinc look in the BOINC-Wiki |
Kim Vater Send message Joined: 27 May 99 Posts: 227 Credit: 22,743,307 RAC: 0 |
Have a look in this thread: http://setiathome.berkeley.edu/forum_thread.php?id=31447 Kiva Greetings from Norway Crunch3er & AK-V8 Inside |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
Well, increasing your RAC will also increase the proportion of science done. This is so often misunderstood here. The more throughput, the more science, the more credit, simple as that! Aloha, Uli |
Saenger Send message Joined: 3 Apr 99 Posts: 2452 Credit: 33,281 RAC: 0 |
Well, increasing your RAC will also increase the proportion of science done. This is so often misunderstood here. The more throughput, the more science, the more credit, simple as that! Depends on the method of increasing. Just fiddling with benchmarks only gives higher claims, with no added value sciencewise. And again: it's not the questio asked. kivas hint was the right one, it seems to be the right thread for this. |
MattDavis Send message Joined: 11 Nov 99 Posts: 919 Credit: 934,161 RAC: 0 |
I appreciate the optimized app advice but I just don't think it is for me. I enjoy the "hands off" approach to BOINC, where applications will update themselves. Some people like to babysit and tweak their computers, and they like optimized apps. I like to keep everything standard - less can go wrong that way, and I don't have to keep an eye on things. ----- |
Steve MacKenzie Send message Joined: 2 Jan 00 Posts: 146 Credit: 6,504,803 RAC: 1 |
Well... again, Matt's question was. Why is same resource share, all stock apps, not getting the same credit over time. To expand on Matt's question, Do the apps that claim more credit get more of their fair share of our cycles. I'm not implying anything ( Safety's on ! ). Steve |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
Depends on the method of increasing. Just fiddling with benchmarks only gives higher claims, with no added value sciencewise.Using an optimized client (proved by the project leader @ Einstein!) produces the very same valid results as the stock cruncher, only faster! To the main question: Einstein had a few issues and downtimes in the last days. This delays the granting of credits and lowers the RAC temporarily, as it is based on credit granted. Aloha, Uli |
MattDavis Send message Joined: 11 Nov 99 Posts: 919 Credit: 934,161 RAC: 0 |
Well, I can add one piece of information. I forgot I had a friend's computer crunching for me. It's only on partially, and only has Seti, Predictor, and LHC on it. http://stats.kwsn.net/host.php?proj=all&hostid[]=393644 Still, that shouldn't cause all the issues. ----- |
Saenger Send message Joined: 3 Apr 99 Posts: 2452 Credit: 33,281 RAC: 0 |
Using an optimized client (proved by the project leader @ Einstein!) produces the very same valid results as the stock cruncher, only faster! Don't mix up Client (BOINC) and Application (Einstein). Optimized apps improve the science, "optimized" clients only "improve" the credit claim. |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
Don't mix up Client (BOINC) and Application (Einstein). Optimized apps improve the science, "optimized" clients only "improve" the credit claim.Just to make sure: I was talking about the science application, i.e. 'the cruncher', not the BOINC-client. Sorry for the confusion. Aloha, Uli |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
Well, I can add one piece of information. The machine running without Einstein would definitely be a part of the difference. Other possible things to look at: Is there a large backlog of pending credits for Einstein? Are a large portion of them failing to validate? Is the machine running in EDF? If so, then the RAC will be high for Einstein later. BOINC WIKI |
Bill Hepburn Send message Joined: 18 May 99 Posts: 11 Credit: 13,149,813 RAC: 12 |
Well, I can add one piece of information. I have noticed the same effect the entire time I have been running BOINC. It also seems that all credits are not necessarily created equal. One hour of CPU time expended on SETI does not result in the same number of credits being requested (or granted) as one on Rosetta, or LHC, or CPDN or whatever. In my experience the difference is as much as a factor of 2. |
Rudy Send message Joined: 23 Jun 99 Posts: 189 Credit: 794,998 RAC: 0 |
I appreciate the optimized app advice but I just don't think it is for me. I enjoy the "hands off" approach to BOINC, where applications will update themselves. While I would'nt recomend Einstein optimized app with only about a week to go there are a few clarifications needed. Akos's optimized applications is for specific processors, and are exact replacements for the stock application. No babysitting required. No App_info required. When a new official app is release it will automatically replace the Akos optimized application which is the same size and name as the stock application. |
W-K 666 Send message Joined: 18 May 99 Posts: 19073 Credit: 40,757,560 RAC: 67 |
The types of computer system can have a fairly big difference in performance when compared across projects. The P3 933 MHz, is good on Seti and poor on Einstein and the Celery 1100 MHz is excellent on Einstein and abysmal on Seti. Throw in mixes of cpu manufacturer and motherboards and only patient observation will tell you were to match a specific computer to specific projects. Or how to set the resourses. And it all can change when the application changes. Paul D. Buck, I hope he's fit and well doing whatever, had two computers that where the same except for the motherboard. The performance difference was staggering. Unfortunately I think that was one page that was not upgraded to the Wiki. ANdy |
Steve Cressman Send message Joined: 6 Jun 02 Posts: 583 Credit: 65,644 RAC: 0 |
The types of computer system can have a fairly big difference in performance when compared across projects. The P3 933 MHz, is good on Seti and poor on Einstein and the Celery 1100 MHz is excellent on Einstein and abysmal on Seti. And that goes along with what I tell people when I help them build systems or upgrade. Your computer can only go as fast as the slowest parts in the system. And since the everthing is connected to the motherboard that is one of the places where you can find some of the biggest differences in performance. :) 98SE XP2500+ @ 2.1 GHz Boinc v5.8.8 And God said"Let there be light."But then the program crashed because he was trying to access the 'light' property of a NULL universe pointer. |
©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.