| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Mapping Cancer Markers Forum
Thread: incapable of running work... |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 8
|
|
| Author |
|
|
Diogo Silva
Cruncher Joined: Apr 1, 2020 Post Count: 3 Status: Offline |
Hello there!
Im using the last version of boinc and since I've updated, it cannot run my usual work (mapping cancer markers). It starts but after around 2 min. it stops and restarts again and again...so im unable to get jobs done. What should I do? Thanks in advance. Best Regards |
||
|
|
bfmorse
Senior Cruncher US Joined: Jul 26, 2009 Post Count: 444 Status: Offline Project Badges: 
|
What operating system is it running under?
Which specific BOINC version have you installed? Does it report an error for the work unit, either in the BOINC log or elsewhere? If so, copy/paste the information here. Someone should be able to help you. If you haven’t already done so, sometimes when you reboot the computer, the problem clears. |
||
|
|
Diogo Silva
Cruncher Joined: Apr 1, 2020 Post Count: 3 Status: Offline |
the version of boinc is 7.20.4
im running on iMac m1 no error...it just restarts after a while for another job without completing... |
||
|
|
bfmorse
Senior Cruncher US Joined: Jul 26, 2009 Post Count: 444 Status: Offline Project Badges:
|
Seems like it should work. but here is link to current versions:
----------------------------------------https://boinc.berkeley.edu/download_all.php [ETA} at the top of that page: Select COMMUNITY -> MESSAGE BOARDS Then enter the search "iMac m1" and click on SEARCH FORUMS and review from there. You do not need to be logged in to read, but you do to post. I know you look forward to Happy Crunching. [Edit 1 times, last edit by bfmorse at Jan 20, 2023 3:05:04 PM] |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges: 
|
----------------------------------------  - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
shanen0
Cruncher Joined: Feb 4, 2021 Post Count: 27 Status: Offline |
Just stopped by to see the status of the same issue. Goes back a LONG time. I only run BOINC on Windows 10 machines and one MacBook Pro, but the Apple machine has been flat-lined for a long time. When it gets OPN tasks, they run to completion and the line moves up, but there haven't been any of them for a while now. The problem has lasted over various updates of the Mac's OS and of BOINC, so this is NOT a transient of any sort.
----------------------------------------I have not noticed any directly comparable problems with Windows machines. (Other problems, however.) Two other symptoms not reported (in this particular thread) involve MCM tasks that do not reset around two minutes. They are rare, but sometimes show up. I have noticed two flavors of this one. The one I first noticed runs to 100% and then stops without finishing. It just stayed in the running status for several more days. (Now I don't remember if it timed out and killed itself or if I finally nuked it.) More recently I noticed one that ran to completion (out of MANY failures). That unit seemed to be completed and reported as normal, but no jump in the points for the machine. I waited a couple of days to see if it was some kind of slow accounting problem, but now I think it was just wasting electricity. Hey, if I want to waste electricity, Bitcoin is always eager for more delusional suckers to join their lotteries. One more thing. The MCM program is obviously buggy. How can anyone trust research based on results from software that is broken? It does matter. [Edit 3 times, last edit by shanen0 at Jan 22, 2023 6:50:34 PM] |
|||
|
||||
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 1320 Status: Offline Project Badges:
|
One more thing. The MCM program is obviously buggy. How can anyone trust research based on results from software that is broken? It does matter. For what it's worth, I think you'll find the issues with MCM are in the control code rather than the science part...I find it interesting that these issues came to the forefront round about the arrival of Monterey O/S version (lots of stuff in the WCG forums late in 2021, such as this Someone raised a ticket on it, but I suspect getting ready for the move to Krembil put that on the back burner... The problems at both ends of an MCM run sound like process timer problems -- the perpetual restarting might be down to processes not getting enough CPU time to complete set-up during the initialization phase, and the issues with tasks sitting around at 100% and never going away could be down to the wrap-up process not completing within a reasonable time but not being detected as such for some reason. Yes, bugs somewhere, but... I think there was a change in the system timer routines around then -- if the MCM1 program for Apple was compiled with static libraries that might explain things... Some people reporting issues on Apple kit were using the "% of CPU" to cut down on fan noise, temperature or whatever -- that wouldn't help if there was a timer issue, and one or two users noted an improvement when they set that at 100% (and limited the number of cores instead?) It didn't resolve the "run forever" issue for all of them, but it did seem to solve the start-up problem. Of course, not having a native version for Apple M1 (and newer) chips doesn't help either, as that introduces Rosetta 2 into the equation for users who have new Macs; it may or may not affect the MCM1 app in a different way. Perhaps once other problems have been resolved, the WCG team will be able to give enough time to the rest of their day job (as in MCM1!) to do something about an Apple silicon version. In conclusion, whilst I agree that there's an issue, I'm not sure that the answer is in the science code :-) I'll continue to run MCM1 (and other projects) until someone proves that any bugs affect the science rather than being irritating (and/or "cosmetic") Cheers - Al. |
||
|
|
shanen0
Cruncher Joined: Feb 4, 2021 Post Count: 27 Status: Offline |
Thanks for your reply and you got me to look at it again. Mixed results to report back...
----------------------------------------Your guess about the CPU limit was correct and seems to have fixed that particular flavor of problem for that particular computer. But I'm pretty sure I had a reason for setting it, though I can't recall what it was now. However, changing such an irrelevant setting should not completely break anything. It should just run slower. I was so encouraged that I decided to restart WCG on one of my largest machines. Which soon triggered the reappearance of one of the worst bugs, a total freeze requiring a hard reset. Yeah, I know correlation is not causation, but I also know that I don't recall ever seeing that crash without WCG being active. Sorry, but I know buggy code when I bang my head on it. I can hope you're right that the scientific results are accurate, but if so, it's in spite of bugs. Capsule history: I started with this kind of stuff many years ago. Back then seti@home may have been the only ballgame in town. I worked my way up to top 1%, mostly because I started using more computers around then. (Currently down to seven and a half, but only three running BOINC.) I ran WCG for some years and passed 4 million points in spite of frequent frustrations. Those problems finally drove me to switch to other projects. I earned 12 million points on one of them before it was completed. Then I came back to WCG. At that time it was still troublesome, and I just want to help out, not fight problems. But I couldn't find a better project (considering objectives of the project and various other criteria), so I hung around and wound up watching the transition away from IBM's sponsorship. Seemed like a long and rather rough transition, but they finally did it. Not sure reproducing all of the problems should be counted as substantial progress. And I definitely feel like any research results based on WCG code need to be carefully validated. Preferably independently. [Edit 2 times, last edit by shanen0 at Jan 30, 2023 5:43:00 PM] |
||
|
|
|