Posts by Toombra

1) Message boards : BOINC Manager : Unfinished Tasks Question. (Message 100163)
Posted 31 Jul 2020 by Profile Toombra
Post:
The deadlines are the only thing BOINC cares about. But within reason, and you as the user has the responsibility to allow BOINC to do all work before the deadline, by pairing projects together with similar deadlines and not filling up the cache to where it's absolutely impossible for it to do all that work before their deadlines.


I just feel like the BOINC should be getting as many WUs done as possible, and if it's told by my preferences screen to Store at least 0.1 days of work and up to an additional 0.1 days of work... then it should be managing it's download requests in a manner to get the optimum amount of WUs done.

Right now, I'm watching as a mt WU has hijacked the system to process, while several other project WUs have now missed their deadline because they were not allocated any CPU to get the work done in the allotted time. (*also a reduction in general uptime due to heatwaves). I watched for a whole 5 days where they could have been completed and then it could have been resource shared to the single mt thread, however the logic dictated different.
2) Message boards : Projects : nanoHUB_at_home has failed every single task -- work time exceeded (Message 99949)
Posted 14 Jul 2020 by Profile Toombra
Post:
Science United has recently decided to start throwing nanHUB_at_home work units at my computer. Unfortunately not a single one of these is processing. All have an estimated time of 3min 47sec, and are timing out after 1h 33min of crunching time. Each work unit has log files similar to the following:

2020-07-10 02:27:29 AM | nanoHUB_at_home | Aborting task 07839998_45_0: exceeded elapsed time limit 5576.75 (86400.00G/15.81G)
2020-07-10 02:27:31 AM | nanoHUB_at_home | Aborting task 07839998_37_0: exceeded elapsed time limit 5576.75 (86400.00G/15.81G)
2020-07-10 02:27:46 AM | nanoHUB_at_home | Computation for task 07839998_45_0 finished
2020-07-10 02:27:46 AM | nanoHUB_at_home | Output file 07839998_45_0_r1958071825_0 for task 07839998_45_0 absent
2020-07-10 02:27:46 AM | nanoHUB_at_home | Computation for task 07839998_37_0 finished
2020-07-10 02:27:46 AM | nanoHUB_at_home | Output file 07839998_37_0_r1960253673_0 for task 07839998_37_0 absent


https://imgur.com/a/fDPBmi4

In-use computing is only able to process about 4 WUs at a time with the rest showing Waiting for Memory as per my preferences.

Specs listed below for reference

2020-07-10 02:40:08 AM |  | Starting BOINC client version 7.16.7 for windows_x86_64
2020-07-10 02:40:08 AM |  | log flags: file_xfer, sched_ops, task
2020-07-10 02:40:08 AM |  | Libraries: libcurl/7.47.1 OpenSSL/1.0.2s zlib/1.2.8
2020-07-10 02:40:09 AM |  | CUDA: NVIDIA GPU 0: GeForce GTX 1650 (driver version 451.48, CUDA version 11.0, compute capability 7.5, 4096MB, 3327MB available, 3037 GFLOPS peak)
2020-07-10 02:40:09 AM |  | OpenCL: NVIDIA GPU 0: GeForce GTX 1650 (driver version 451.48, device version OpenCL 1.2 CUDA, 4096MB, 3327MB available, 3037 GFLOPS peak)
2020-07-10 02:40:09 AM |  | Windows processor group 0: 12 processors
2020-07-10 02:40:09 AM |  | Processor: 12 AuthenticAMD AMD Ryzen 5 2600 Six-Core Processor [Family 23 Model 8 Stepping 2]
2020-07-10 02:40:09 AM |  | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni ssse3 fma cx16 sse4_1 sse4_2 movebe popcnt aes f16c rdrandsyscall nx lm avx avx2 svm sse4a osvw skinit wdt tce topx page1gb rdtscp fsgsbase bmi1 smep bmi2
2020-07-10 02:40:09 AM |  | OS: Microsoft Windows 10: Core x64 Edition, (10.00.18363.00)
2020-07-10 02:40:09 AM |  | Memory: 15.93 GB physical, 21.18 GB virtual
2020-07-10 02:40:09 AM |  | Disk: 931.51 GB total, 587.71 GB free
2020-07-10 02:40:09 AM |  | Local time is UTC -6 hours
2020-07-10 02:40:09 AM |  | No WSL found.
2020-07-10 02:40:09 AM |  | VirtualBox version: 6.0.22
2020-07-10 02:40:09 AM |  | General prefs: from https://scienceunited.org/ (last modified 26-Jun-2020 02:52:09)
2020-07-10 02:40:09 AM |  | Host location: none
2020-07-10 02:40:09 AM |  | General prefs: using your defaults
2020-07-10 02:40:09 AM |  | Reading preferences override file
2020-07-10 02:40:09 AM |  | Preferences:
2020-07-10 02:40:09 AM |  | max memory usage when active: 8974.23 MB
2020-07-10 02:40:09 AM |  | max memory usage when idle: 14685.10 MB
2020-07-10 02:40:09 AM |  | max disk usage: 592.37 GB
2020-07-10 02:40:09 AM |  | max CPUs used: 10
2020-07-10 02:40:09 AM |  | don't use GPU while active
2020-07-10 02:40:09 AM |  | suspend work if non-BOINC CPU load exceeds 35%
2020-07-10 02:40:09 AM |  | (to change preferences, visit a project web site or select Preferences in the Manager)
2020-07-10 02:40:09 AM |  | Setting up project and slot directories
2020-07-10 02:40:09 AM |  | Checking active tasks
2020-07-10 02:40:09 AM |  | Using account manager Science United


Since I've written the above post, I've had more WUs downloaded with the exact same issue. I'm manually aborting 100s of these as they'll just waste processing.
3) Message boards : BOINC Manager : Unfinished Tasks Question. (Message 99420)
Posted 24 Jun 2020 by Profile Toombra
Post:
The other problem I'm seeing is that multi-core tasks will "hog" all the resources and instead of being able to get more WU done in general, a couple of multi-core WUs (often with deadlines that are later in my case) will take the focus.

Here is a screenshot that I take tonight, since I will be at work for a majority of the next day, and possibly out for the rest of that evening. Several of these Rosetta WUs are in danger of not being completed, because the scheduler would like to leave them "until the last minute" while it prioritizes the other and then panic and prioritize them to try to get done. You can see that the multi-core Milkyway@Home WUs are even preventing the non-multicore WUs from finishing up and being completed.

https://imgur.com/a/2c3rXwo

I think the deadlines should become more of a priority in the scheduler. Either give more leeway so that panic time is activated sooner or just change the logic so that the scheduler goes "I should do this task that's due in 3 days rather than this task that's due in 2 weeks". It seems to me there's probably a lot of incomplete WUs in the pipeline among those who aren't running 24/7 machines or building systems around BOINC.
4) Message boards : Questions and problems : sigh... 20+ hours of work thrown out at shutdown (Message 42750)
Posted 26 Feb 2012 by Profile Toombra
Post:
sigh, basically annoying but I'm not sure if there's anything anyone can do about this due to the way windows operates.

I run various BOINC projects, among them is the climatepredictation.net . Now climateprediction as many know has really large workunits, the kind that take hundreds of hours to process and travel back to the server with trickle updates. Unfortunately because the length of time is so long to process these, it's really unfortunate when something occurs to disrupt it, basically negating the data.

So what's happening is so silly it makes me angry. Unlike a lot of BOINCers, I turn my computer off from time to time, generally when I'm away for awhile. Now windows during a shutdown always acts the same way; it tries to close down each program and gets confused when something's preventing it from turning off. One such thing is a program such as microsoft word or even notepad asking "Do you want to save your work?". No problem with that, most users close out of the screen, save (or discard) and then try again. Unfortunately when BOINC is set in always run mode.... this point where windows is hung with the dialog box..... results in computational errors with the tasks that try to run. BOINC will keep trying to work on them, and when it can't... will throw up it's hands and give up completely on the workunit. What's especially annoying is that this occurs within about 1 or 2 minutes... without much time for the user to react and fix the problem.

Taking a look at my host page: http://climateapps2.oerc.ox.ac.uk/cpdnboinc/results.php?hostid=1112451 You can see the 3 climateprediction.net tasks from February 24 that are showing ERROR WHILE COMPUTING. These are all because of the windows shutdown situation mentioned above. Each task had approximately 17-20 hours processed already. Gone. all gone.

Really annoying. I'll post this to CPDN's board too...
5) Message boards : Questions and problems : I can always tell when Einstein@Home is running, sluggish (Message 40534)
Posted 4 Oct 2011 by Profile Toombra
Post:

I haven't been paying too much attention between the apps beyond identifying them to each project so it's possible that only one in particular is noticeable while the other isn't. I'll keep a better eye on these going forward to try to isolate performance.

At the moment the Gamma Ray Pulsar Search #1 0.23 is running and using 265MB of memory per process. I don't have an accurate baseline at this exact moment for performance. As I said, I'll try to keep a better eye on it.


The "Binary Radio Pulsar Search (Arecibo) 1.00 (BRP3SSE)" application seems to be one of the ones that's detrimental on system performance. I notice more sluggishness while running 1 instance of this application than I do running 4 of the "Gravitational Wave S6 GC search 1.01 (SSE2)" applications.
6) Message boards : Questions and problems : I can always tell when Einstein@Home is running, sluggish (Message 39997)
Posted 3 Sep 2011 by Profile Toombra
Post:

Though Dagorath didn't see your (edited in?) comment on GPU using, he might be right in suggesting memory contention problems, only for main memory. Since you didn't tell how much memory your host has built in and your hosts are hidden at Einstein, I can't tell for sure if that might be the case.


Sorry, you may be right that memory is the issue. This computer currently has only 4GB of DDR3 ram. so far it's been enough for most things I do and as mentioned I don't notice sluggishness in other boinc apps.

In addition, there are three different applications at Einstein@home (two for CPUs) that use different amounts of memory.


I haven't been paying too much attention between the apps beyond identifying them to each project so it's possible that only one in particular is noticeable while the other isn't. I'll keep a better eye on these going forward to try to isolate performance.

At the moment the Gamma Ray Pulsar Search #1 0.23 is running and using 265MB of memory per process. I don't have an accurate baseline at this exact moment for performance. As I said, I'll try to keep a better eye on it.
7) Message boards : Questions and problems : I can always tell when Einstein@Home is running, sluggish (Message 39996)
Posted 3 Sep 2011 by Profile Toombra
Post:
Do you crunch Einstein GPU units? If you do then that could be why games and video playback are sluggish. Tasks running on the GPU don't run "in the background" or at low priority. There is no such thing as "background" execution and low priority on GPUs. With an Einstein GPU task plus a game or video you might be expecting more from your GPU than what it is capable of. Does it have enough memory to run a game/video plus an Einstein task? What model video card do you use? How much memory does it have? How many shaders?

Einstein code is their responsibility but YOU have a responsibility to know what your hardware is capable of and when you are expecting more than it can possibly deliver. My GTX 570 can handle an Einstein GPU task plus video playback but it's a new, fast card with lots of shaders and 1 gig memory.


As I said in my initial post: "P.S. None of these projects use my GPU (incompatible model/brand) so it's not that. Einstein@Home seems to only like NVidia and I happened to buy ATI. oops."

AFAIK it is entirely CPU scheduling.
8) Message boards : Questions and problems : I can always tell when Einstein@Home is running, sluggish (Message 39986)
Posted 3 Sep 2011 by Profile Toombra
Post:
Apologies if this has been posted elsewhere I searched through other threads and didn't see what I was looking for.

My issue isn't an error or anything. I've been a BOINC user for several years and currently am running 6 projects on my year-old computer of 3.4ghz X4 quad-core processor. The projects I'm running are: boincsimap, climateprediction.net, Einstein@Home, Milkyway@Home, orbit@home and SETI@Home. I have also run the World Community Grid project in the past and on other machines. BOINC manager is currently 6.12.33 x64

These projects all behave well for the most part and while I use my computer for many different things, I'm quite content to leave BOINC in "Run Always" mode. It seems to be fairly efficient (with the occasional bug that doesn't occur often) at managing the projects in the background and ensuring they don't interfere with any higher priority processes (which should be every other application) that wants to run. In fact I play a variety of 3D games and the hit to the performance and framerate by BOINC is negligable...

...Except for one project: Einstein@Home. Whenever this project happens to be running, and especially when multiple processes of it are running (the X4 is a multicore chip; 4 "CPUs"), I can tell because of the overall sluggishness that creeps into the computer usage. It's not major, but if I'm wanting to play a game or even watch a video... I can tell when it's running and often suspend BOINC operations for this period (on a related note; being able to assign a hotkey would be nice).

While I understand the code of the Einstein@Home application is ultimately the responsibility of the project and it's programmers. I thought I'd take a moment to highlight that this project does not seem to behave as well in the "background" as the other projects I have experienced, even while it it uses up approximately the same amount of CPU power and memory (and definitely less than Climate@Home).

P.S. None of these projects use my GPU (incompatible model/brand) so it's not that. Einstein@Home seems to only like NVidia and I happened to buy ATI. oops.
9) Message boards : Questions and problems : BOINC manager does not switch applications (Message 37051)
Posted 2 Mar 2011 by Profile Toombra
Post:
Same problem here. Actually it wasn't a problem until the last day or two and I'm still not sure the cause since I haven't done any manually suspending the work.

Having said that, I really don't like the behavior. If the user does suspend the processing for whatever reason... then that's a user decision, not the project's and shouldn't affect the debt calculation...
10) Message boards : Questions and problems : Boinc installer problem with windows .msi (Message 24410)
Posted 19 Apr 2009 by Profile Toombra
Post:
Thanks Ageless. I didn't come across that document in any of my searches. I'll give it's instructions a shot!
11) Message boards : Questions and problems : Boinc installer problem with windows .msi (Message 24408)
Posted 19 Apr 2009 by Profile Toombra
Post:
Hi, the other day I decided that I wanted to upgrade my BOINC on one of my computers since I was a version or two behind. Unfortunately after I uninstalled the old version, I was unable to install the new version. Specifically the error I was getting was that part-way through the installation, the program decided that a "network resource was unavailable" and that it could not find "boinc.msi" even though that's the file it was currently running from.

Screenshot link - http://img23.imageshack.us/gal.php?g=boincinstallerror1.jpg

I went back to download an older version and discovered that I had the same issues with the installer. I'm sure this has something to do with my system specifically or msiexec, however I can't seem to figure it out. I've verified that I'm able to run other msi installers that I have correctly. Other applications and games are able to install just fine, so it's only BOINC.msi causing issues. I've also upgraded msiexec to the latest version XP can handle after tracking down the necessary installer on Microsoft's website.

Any help would be appreciated. I've already emailed Rom Walton as per the "installation problems" thread but have not heard anything back.

Right now my CPU is being wasted whenever I have to leave it on and aren't using up all it's juice. I don't like that.
12) Message boards : Questions and problems : Problems with project executables after System Hibernation (Message 22861)
Posted 3 Feb 2009 by Profile Toombra
Post:
BOINC Manager: 6.4.5
Platform: WindowsXP/32bit
SETI@Home: setiathome_6.03_windows_intelx86.exe
SIMAP: simap_5.10_windows_intelx86.exe
Lieden Classical: classical_5.53_windows_intelx86.exe

I've been having consistent problems over the past month or so, since I updated to this current version of BOINC manager on laptop. The settings on this laptop that I use are a bit sub-optimal for BOINC in order to not create too much heat (especially when I'm actually placing it on my lap). Specifically the settings are to do only work after about 45 seconds of idle time. This seems to work rather well, however problems arise when I'm done using my laptop and put it into HIBERNATE mode. Upon exiting HIBERNATE mode, a couple of project applications jump to 100% cpu usage at all times and often return "<TASK> exited with zero status but no 'finished' file" error messages during this period. Killing the running application in process manager makes it return to normal settings.

Not sure what's causing this bug but it seems to have appeared either with the 6.4.5 version of the manager or the specified versions of each application.
13) Message boards : Questions and problems : 6.4.5 wants to make me work for it... (Message 22171)
Posted 29 Dec 2008 by Profile Toombra
Post:
It also seems that 3) is a normal crash/wayward process with the particular project exe. I haven't had one of those for a long time so I just didn't recognize it. Even so, it badly impacts the system and it would be nice if BOINC was able to detect and terminate this. I know plenty of users who would end up complaining "BOINC screws up my computer" and go uninstalling it (assuming they managed to identify that as the problem and not spend hours fiddling with other settings). Watching the process manager, all it seemed to be doing was trying to launch new versions of the same EXE and then wondering why they weren't getting any CPU time because the old version of the EXE was still sucking down 99% while being suspended in the manager.

Could be the Leiden Classical project executable isn't as stable as others. Even so, the framework needs to deal with this better.

6.4.5 was brought out for people who wanted to run cuda apps. But a lot of problems from 6.2.* weren't fixed in the release, and a few new ones seem to have appeared, especially with cuda.

Actually reading through Ageless' version posts it seemed like more had been fixed/changed (plus it did go up 2 numbers :P... though that's a rough guide at best). But, you may be right and I shouldn't be dealing with this version... on the other hand, I just installed 2 other BOINCs of 2.4.5 on family members' PCs over the hollidays so I'm not crazy about having to think up mismatched versions.
14) Message boards : Questions and problems : 6.4.5 wants to make me work for it... (Message 22168)
Posted 29 Dec 2008 by Profile Toombra
Post:
Hi everyone, I've been a steady user of BOINC on my home Windows (all 32 bit) machines for a long while now, casually upgrading through the versions as they became available AND contained features that seemed worth a new install. For reference, I currently have 4 personal computers signed up on my account: 1 XP SP3 machine, 2 XP SP1 machines and 1 Vista SP1 machine, none of which have the new Nvidia 3D crunching unfortunately, due to ATI cards and embedded chipsets. Most of the time it's been fairly painless but upgrading from 6.2.x to 6.4.5 seems to have caused me a couple of issues.:

1) First, and I've seen this referenced in other posts for older versions, but this latest version doesn't seem to want to run as boinc_users at all. From a new installer from Administrator to manually adding the group to user accounts... it just will not let BOINC Manager connect to the running process/service. The only success I've had is by adding all users as boinc_admin group, which while less than perfect is acceptable for my setup.

2) BOINC doesn't seem to be listening to my settings override as well as previous versions. Specifically I'm referring to the CPU Usage setting and the "only do work when computer has been idle for" setting. It's hard to describe, but I see this version leaping to 100% usage on the Task Manager or the Temperature monitor when it shouldn't be.

3) On ONE computer and only ONE... I've been working away on something and had the BOINC (specifically Leiden Classical) jump to 100% usage and slow everything else done. It doesn't seem to be runnning at lowest priority if my application performance is any indication. Twenty minutes ago (before searching forums) I told it to suspend and while BOINC Manager acknowledges this in it's log, the project's EXE is still burning away at 100%.

Any help on these issues would be appreciated... or Suggestions? Comments? Laughter?

Here's hoping for 6.5, 6.4.6 or 6.4.5.1 ;)




Copyright © 2021 University of California. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.