|
1)
Message boards :
Number crunching :
WUs stuck on 100% completion and then had to abort (ANDROID)
(Message 4391)
Posted 14 Jul 2020 by James W Post: I've seen this issue more of late, especially on my RCA (Alco) tablet with 4 processors. I have 4 Android devices working on Universe@Home, as these don't have enough RAM to work on medical projects fighting COVID. The process times can vary greatly, between 130K seconds to over 606K seconds. If I aborted every task that seemed to get "stuck" at 100% I wouldn't get much done. I note that even though the task shows at 100% that the clock is still advancing, so I assume it's still crunching. Occassionally on my RCA tablet I'll suspend the task, which starts another new task. I'll then unsuspend the 100% task and most times it will start back at 0%. The task is NOT lost, as the clock will continue where it left off. When jobs (Universe BHspin) complete on computers, in the BOINC manager I note there are a number of sections to each task. I've concluded from this that the same process is going on with the Android jobs and that when it "starts over" it is actually starting another section of the task. Krzysztof would need to verify that for us, however. In other words, I'd suggest to NOT abort these jobs, but try to coax it to continue on as described above or by other means. My longest job so far has taken 7 days (606K seconds), so don't panic. |
|
2)
Message boards :
Number crunching :
Upload failure: file size too big
(Message 4166)
Posted 19 Apr 2020 by James W Post: <edit>See thread "WU Error -> WU error -> WU error" and Message 4161 from Project Administrator, where he explains problem. Though he boosted WU size limit to 2G, there are still larger files generated. The task reported below has Peak disk usage of 2.31 GB. WU=44121708; Task=96754296 Application: Universe ULX v0.12 windows_x86_64 Client state: Compute error Exit status: 0 (0x00000000) Stderr output: upload failure: <file_xfer_error>Most of these "ulx" WUs crunching on my Windows machine have failed due to the file size being too big, or other language to same affect. In this WU alone, mine was only one of six failures, with one being an aborted task. Other WUs have similar errors, such as "Exit status: 196 (0x000000C4) EXIT_DISK_LIMIT_EXCEEDED." Appears this type of Task/WU needs to be reviewed. No wonder someone has resorted to aborting this type of task. |
|
3)
Message boards :
Number crunching :
Why is Project requesting new work units every hour??
(Message 3788)
Posted 9 Sep 2019 by James W Post: This last week or so I've noticed the "project" has been requesting new work every hour, irregardless of how many jobs already are in the cache. Is there a reason for this?? BOINC Manager event log shows: 9/9/2019 1:54:36 AM | Universe@Home | Sending scheduler request: Requested by project. |
|
4)
Message boards :
Number crunching :
Some task run longer than usual
(Message 3787)
Posted 9 Sep 2019 by James W Post: But many have forgotten that there was no storm of indignation when about a year ago (or two) the credits were massively increased and so virtually all the hard-won work by new entrants was devalued. That was the normal credit level for years. Only as a reminder. Amen to above! I'm here for the science and to explore the universe, not a credit race. . . . and I'm not into Bitcoins, etc. |
|
5)
Message boards :
Number crunching :
Run times versus CPU times
(Message 3348)
Posted 12 Feb 2019 by James W Post: Thanks for the response. However the host in question doesn't have multithreading, just 4 processors. |
|
6)
Message boards :
Number crunching :
Run times versus CPU times
(Message 3343)
Posted 11 Feb 2019 by James W Post: I notice today that once again CPU run times are taking longer than total run times for my PC jobs the last couple days. How is this possible? How could the CPU run LONGER than the job as a whole. A puzzle. Misplaced coding for data chart/spreadsheet? |
|
7)
Message boards :
Number crunching :
Zero credit for valid Work Unit
(Message 3284)
Posted 5 Feb 2019 by James W Post: I was surprised to see Work Unit 22960237 with zero credit, despite it being "completed and valid." Other tasks completed today had credit, including some scored using the revised credit system with roughly 1/10 of the previous 667 across the board. I can deal with the revised system as long as it is fair for the time and energy used by my hosts. Hopefully the zero credit is just a glitch in the new system and can be corrected. Thank you. |
|
8)
Message boards :
Number crunching :
What happend 29 Dec till now?
(Message 3189)
Posted 4 Jan 2019 by James W Post: Server re-sent a part of my workload which turned some completed tasks in invalid results. Not nice. Same thing happened to me with task now past deadline, though luckily only lost one task (47444475), which was resent and completed before I realized server was back on line. My BOINC Manager had "backed off" on transfers because server was down, and wasn't scheduled to resend for many hours yet. I therefore had to manually update Universe, but too late for task noted above. |
|
9)
Message boards :
News :
Server crash
(Message 3098)
Posted 21 Nov 2018 by James W Post: Work Unit=19398066 Task=44010287 Status: Timed out - no response Report Deadline: 20 Nov 2018, 9:59:29 UTC 11/18/2018 1:06:52 AM | Universe@Home | Computation for task universe_bh2_180328_246_718754282_20000_1-999999_755400_0 finished However, this was one of several WUs which were unable to report until 10:35 a.m. local time (USA-PST). This is apparently one of WUs you spoke about which were not backed up due to drive failures, so server treated as an unreturned WU? The results were uploaded 11/18 about 1 a.m. as noted above. This info lost as well? Only job I found out of the 7 reported this morning which timed out I've noted so far. FYI. I don't envy your task of getting things back into working order. Take care! |
|
10)
Message boards :
Number crunching :
Task Website pages with error message
(Message 3066)
Posted 28 Oct 2018 by James W Post: Since the BOINC server upgrade, I've noticed the following error message on all my Task Website pages. Otherwise data is uneffected. Appears to have something to do with formatting? Just an FYI. Thanks. Notice: Use of undefined constant ALIGN_RIGHT - assumed 'ALIGN_RIGHT' in /home/boincadm/projects/universe/html/inc/result.inc on line 404 |
|
11)
Message boards :
Number crunching :
BHspin v2 v0.01 (android_arm_pie) restarts after at 100% completed
(Message 2913)
Posted 27 May 2018 by James W Post: As these apps seem to have a number of segments to them, is this "restarting" action related to completing each of these parts? |
|
12)
Message boards :
Number crunching :
BHspin v2 v0.01 (android_arm_pie) restarts after at 100% completed
(Message 2897)
Posted 22 May 2018 by James W Post: WU=16013977; Task=36075357 CPU Type: ARM -- ARMv7 Processor rev 0 (v7l) OS: Android 3.10.49-9411209 (Android 6.0.1) This is my 2nd WU that seems stuck at 100% and still running and with clock still ticking, currently at 38K seconds runtime. Just finished a WU with this same issue on same device which reached 100% and then started over again 4 times before finally coming to completion, using a total of 435,553.24 sec run time before completion. I hope this won't be a problem with my Android_arm WUs going forward. Stderr output in part for 1st problem job shows the following multiple times: WARNING: linker: ../../projects/universeathome.pl_universe/BHspin2_1_arm-android-linux-gnu has text relocations. This is wasting memory and prevents security hardening. Please fix. After this last line, job finished successfully! |
|
13)
Message boards :
Number crunching :
How do you earn credits?
(Message 2417)
Posted 4 Oct 2017 by James W Post: Well, #1 this project is not "seti" (Search for extraterrestrial intelligence). I would consider it more as astronomy -- mapping the universe. I assume you've installed the BOINC application/program? After you've joined the project, which I see you just did today, you download work units (WUs). Once your computer/device has correctly computed these and returned them to the project, if no errors/issues you would get the allotted credits. (Simplified explanation) |
|
14)
Message boards :
Number crunching :
Cannot Upload or Download Tasks
(Message 2404)
Posted 30 Sep 2017 by James W Post: For those having upload/download issues, please see the posting in News board titled "Server task generating breake..." This explains what Krzystof plans to do to find the cause of this issue which some hosts are experiencing. |
|
15)
Message boards :
Number crunching :
Log-in Problem
(Message 2397)
Posted 29 Sep 2017 by James W Post: Do you experience this problem from server-moving date? Had no problem with logging in after the move the end of Aug. However, began having this problem about 2 weeks or so ago. Able to log in using the authenticator code, however. |
|
16)
Message boards :
Number crunching :
Log-in Problem
(Message 2379)
Posted 26 Sep 2017 by James W Post: Over the last couple weeks I haven't been able to log in with Email & password, which I've double checked as correct in my acct. All that comes up is window titled "Unable to Handle Request" and request to use correct Email and password. Thankfully, I was able to get in using the authenticator code. Am I the only one with this problem? I replaced password, which didn't fix problem. Thanks for any ideas. |
(English)
(Polski)