Granted credits
Message boards :
Number crunching :
Granted credits
Message board moderation
Previous · 1 · 2
| Author | Message |
|---|---|
|
Send message Joined: 9 Jun 10 Posts: 19 Credit: 2,233,932,323 RAC: 0 Level  Scientific publications  
|
Hello, I have a similar problem (or rather anomaly) - my credit got quadrupled today from ~70k to ~280k. I noticed that I finished a long GPU task, but why is this one particular task worth so much? How to find out? I am curious too. I finally got some work done but the credits granted is surprisingly high: https://www.gpugrid.net/results.php?hostid=521591 That's 40-50 cr/sec... PS: Your hosts are actually hidden 
|
 Send message Joined: 12 Jul 17 Posts: 404 Credit: 17,408,899,587 RAC: 0 Level  Scientific publications  
|
For the last 79 valid WUs today I calculate: Mean, Standard Deviation 57.12, 16.95 Credit/RT-second 60.05, 16.98 Credit/CPU-second The standard deviation is almost 30% of the mean. Lots of variability driven by outliers both high and low, e.g. 112, 120, 26.18, 22.47 |
|
Send message Joined: 9 Dec 08 Posts: 1006 Credit: 5,068,599 RAC: 0 Level  Scientific publications   |
Hello, I have a similar problem (or rather anomaly) - my credit got quadrupled today from ~70k to ~280k. I noticed that I finished a long GPU task, but why is this one particular task worth so much? How to find out? It was actually a mistake in the MDADeq series. Will be fixed in the next series. |
|
Send message Joined: 16 Jun 12 Posts: 17 Credit: 292,288,806 RAC: 0 Level  Scientific publications
|
My credit per unit dropped 75% yesterday. I will see about today as I get a few units to crunch. |
|
Send message Joined: 17 Sep 18 Posts: 11 Credit: 1,857,385,729 RAC: 0 Level  Scientific publications
|
250% increase in run times, for a quarter of the credits. It is like pre-Turing days :( |
 Send message Joined: 13 Dec 17 Posts: 1419 Credit: 9,119,446,190 RAC: 731 Level  Scientific publications 
|
My credit per unit dropped 75% yesterday. I will see about today as I get a few units to crunch. Toni stated in this thread the granted credits for the new batch of work was miscalculated and would be corrected. https://www.gpugrid.net/forum_thread.php?id=4826&nowrap=true#53557 Now the runtimes and granted credit are more in line with the previous work from the end of last year. The ultra high credit lately was just a fluke and not something to expect on a regular basis. |
|
Send message Joined: 9 Dec 08 Posts: 1006 Credit: 5,068,599 RAC: 0 Level Scientific publications |
Exactly. To clarify, I won't change already-assigned credits, just future WUs. Credits are proportional to the computations performed (approximately linear in simulation steps x number of atoms) according to a ratio we have been trying to keep constant since many years. |
|
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 351 Level Scientific publications
|
Toni stated in this thread the granted credits for the new batch of work was miscalculated and would be corrected. Just treat it as your Christmas Bonus. Now back to work as normal! |
|
Send message Joined: 16 Jun 12 Posts: 17 Credit: 292,288,806 RAC: 0 Level Scientific publications
|
I hope you also had a nice Christmas bonus, lol. |
|
Send message Joined: 17 Sep 18 Posts: 11 Credit: 1,857,385,729 RAC: 0 Level Scientific publications
|
Thank you for the explanation. Back to the grindstone it is then :-) |
©2025 Universitat Pompeu Fabra