| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Oddity in BOINC Credit |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 26
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I replaced two old machines at work last week with new units and kept the existing network names. I saved the BOINC directory and just copied it back over onto the new machine as well after installing BOINC, and it seemed to work fine, the workunits fired up, I ran new benchmarks, everything was good. However I noticed a few days later that the BOINC credit claimed was ridiculously low (12 points for Cancer units when a clone unit that I installed BOINC on fresh is getting 50-60 per unit). This happened on both units. Thinking something was wrong, I reset the project, it made no difference. I uninstalled BOINC completely from one, deleted the BOINC directory, wiped out the registry entries, and reinstalled, and its still the same story. However I have clone units that I installed from scratch on and they're all claiming 50-60 per unit. I have no idea why this is happening, but it makes me believe that something is wrong somewhere. I'm not in it for the BOINC credit, but, the disparity makes me wonder what I did wrong and why a reinstall didn't fix it.
Another thing I noticed, on the two new ones I copied over, they only download enough work to work on at that moment, whereas the other units that are of similar power (and requesting BOINC credit more accurately) download a large queue of work at all times. And yes, they're using the exact same device profiles. Any ideas? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Claims are rather meaningless these days. What is being *granted*?
If a computer is routinely over-claiming, then if you switch that with a computer that under-claims, WCG will take a while to adjust. WCG have a new mechanism that will NOT create a new device if you completely wipe and reinstall. The moral of the story: don't reinstall to fix credit issues. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for the quick reply. What I'm seeing is my new machines are claiming between 10-15 points for each work unit, and the other result from another computer for the same work unit claims anywhere between 50-100. However we're only being granted what I claimed (10-15). This is for all of them. So essentially, whoever is getting sent the same work units I am on these new machines, is getting the super low credit I am as well, but they're claiming much more.
Does that make sense?  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
What are the CXPU benchmarks for these 'new' mchines?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
1/21/2008 10:29:36 AM|World Community Grid|Finished upload of X0000041950507200411121958_0_0
1/21/2008 10:38:55 AM||Running CPU benchmarks 1/21/2008 10:38:55 AM||Suspending computation - running CPU benchmarks 1/21/2008 10:39:27 AM||Benchmark results: 1/21/2008 10:39:27 AM|| Number of CPUs: 2 1/21/2008 10:39:27 AM|| 2620 floating point MIPS (Whetstone) per CPU 1/21/2008 10:39:27 AM|| 5869 integer MIPS (Dhrystone) per CPU 1/21/2008 10:39:28 AM||Resuming computation |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Strangely enough, one of the computers in question is now requesting around 40 BOINC credit points per work unit. The other is still requesting around 13. I guess maybe this is correcting itself.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
That's what ought to happen. Report back if the other one doesn't normalise soon.
I'm not sure what the best way of avoiding this is. Generally it is a good thing when a computer keeps the same host id. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Well the one has completely corrected its claimed BOINC credit, its right where the other clone units should be. The other one hasn't changed, at all. Its still requesting meager 12 and 13 BOINC credits for the same type of work the other units are getting 50-60 for. It seems like whatever mechanism that's supposed to normalize this hasn't worked for this one. Any ideas on what I can do?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
A few questions:
Are the benchmarks above for this machine or the other? Are you running HCC exclusively? If so, have you checked in Task Manager whether the page fault delta is unusually high? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello billroland,
I can't think of anything off the top of my head. So . . . that means I need more information to see if that gives me an idea. Could you copy the message tab in BOINC from boot through download and sending back a completed work unit? Then add the information from your Results Status page for that work unit. Lawrence |
|||
|
||||
|
|