1) Message boards : Number crunching : Updated GPU Requirements (Currently not supporting GPU tasks) (Message 64194) Posted 19 Dec 2015 by disturber Post: I have both the 7970 and the R9 280x cards. Both are in computers with PCI-e v3 slots. There are differences as to the cpus and windows versions. But I found that if I set the core clock and memory clock the exact same, I cannot distinguish the speed between them. Both cards are capable of 41-42 sec running two tasks concurrently, as long as both tasks are not synchronize. So each day I check and make sure this is the case, otherwise when the tasks are synchronized, computation time increases to 44-45sec. Performance of gpus is easy to check by looking at the reviews of AMD cards on Anandtech. They publish the specs of the chips in each card. The spec to look for is the SP to DP ratio. There are not many AMD cards that have a ration of 1/4. The graphic card vendors have been crippling their chips for a while now, as they make the high DP ratio available in the multi thousand $ cards. The nVidia cards have pretty pathetic SP/DP rations except for Titan cards, but those are $1000 or higher. After I learned of this information last year, I went to look for refurbished or used 7970/280x cards. People were dumping them because they no longer made any money mining bitcoins. My local Microcenter store had several refurbished ones at less then $180, so I picked them up. You may still be able to find them on ebay for a good price, but without the return policy I had at the local microcenter store.

2) Message boards : Number crunching : why so many validation inconclusive? (Message 64193) Posted 19 Dec 2015 by disturber Post: A bunch of these. http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=1028389198 http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=1027234127 http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=1028211808 I guess this is the drawback of having fast cards: http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=1028211808 349 validation inconclusive for my two computers.

3) Message boards : Number crunching : Problem with running two milkyway work units (Message 64168) Posted 12 Dec 2015 by disturber Post: I know this is not quite on topic for my original post. Ever since I started running Einstein cpu work, I have been getting a lot of invalids and validation errors. I don't know if the interaction between cpu and gpu causes that, but I am quitting running any cpu jobs. The amount of credit from that work does not compensate for the loss from the errors and invalids. In addition it does not seem to appreciably effect the synchronization. Has anyone else found that cpu work effects error rate?

4) Message boards : Number crunching : Problem with running two milkyway work units (Message 64159) Posted 6 Dec 2015 by disturber Post: I started running Einstein cpu workunits. It helps some but still they seem to synchronize over time. I stopped running the separation modfits because they are way too short relative to their load time. The Milkyway units are way too short too IMO. No other project has wu that run only tens of seconds. I have not researched it, but does anyone know why this is, or is it because the 280x is just such a fast card for these jobs?

5) Message boards : Number crunching : Computational error on new build (Message 64156) Posted 6 Dec 2015 by disturber Post: I had issues with Windows 10 forcing the new AMD drivers down my throat. I ended up using third party software to remove the new driver from the hidden windows driver repository. After that, it still replaced the driver a few hours later, even after checking all the boxes telling windows not to update drivers. I disabled the service that installs drivers and had no problems since then. In my case the new driver added about 10% longer to the compute time, which is an unacceptable waste of electricity. Mickey, how does the server side software fix this problem? I had the same slower compute time with Einstain@Home, so I rolled back drivers in Windows 7 for that. I think they have newer software, but I could be wrong.

6) Message boards : Number crunching : why so many validation inconclusive? (Message 64155) Posted 6 Dec 2015 by disturber Post: The last few days I have a total of 175 validation inconclusives on both of my machines. Does anyone know what's going? My wingman appears to have the same validation inconclusive. Are these bad batches?

7) Message boards : Number crunching : Problem with running two milkyway work units (Message 64132) Posted 29 Nov 2015 by disturber Post: I have been running two workunits per gpu on two different computers, one running windows 7 and using a R9 280x and the other running windows 10 and using a 7970 card. To get higher production I overlap the two tasks 50% by briefly pausing one of the tasks at the midway point, and then resuming it when it starts calculating a new one. The issue I have is that after a period of time, both workunits synchronize again, increasing my compute time from 41-42s to 44-45s. It may not seem much but that is almost a 10% increase. It is a hassle to to check on this daily and I may give up in that yet. Does anyone have an idea why this would be happening? As you can see, this is not related to either the OS or the card. And even with different cpus and different OS, if they are clocked the same, will produce identical compute times.

8) Message boards : Number crunching : The increase in the calculation time at Win 10 (x64) (Message 64131) Posted 29 Nov 2015 by disturber Post: I had a big problem keeping Windows 10 from installing new drivers. Even after uninstalling and deleting the new driver from the hidden windows store repository, it still downloaded it and re-installed it a few hours later. This happened after I checked all the check boxes disabling driver install with instructions I found on the windows help page. So my solution was to disable the Driver Install service using the administrative program. The 13.251 driver has now been updated for a week after this happened. For my 7970 card, any of the newer drivers slow it down. This applies not only to this project but for Einstaein@Home as well.

©2024 Astroinformatics Group