Posts by Nick Name

1) Message boards : Questions and problems : Errors Moving Data Folder (Message 108332)
Posted 7 Jun 2022 by Nick Name
Post:
Really? Then I don't get why the moving instructions have been provided, if there's another, maybe simpler, way. 'Course especially if those instructions don't work, as the case seems to be. But anyway, huh. That's promising. Maybe there's finally a solution, except that as I said, I use macOS. So I would need the instructions for a script for that OS, instead of one for "Windoze." Any idea what those might be?

Actually I mixed this up. boinccmd is a utility for managing tasks and projects without using the standard GUI manager, so it's not actually needed here. I should have posted this link to the command line options instead (sorry about that). You have to start BOINC and tell it the data location, and the port if you don't want to use the default 31416.

"c:\Program Files\BOINC\boinc.exe"

This is just starting the client. I don't know what this would be on Mac. Maybe all you need to do is type boinc, which I think is what I do on my Linux machines.

 --gui_rpc_port 31450

Don't worry about this if you aren't using multiple clients.

 --dir O:\GPU\AMD\AMD31450

This would be your folder location.

On Linux I use something like this. My machines are offline for the summer so I'm going by memory.

boinc --dir ./Documents/BOINC/CPU/


I intentionally run my clients in unprotected locations so I don't have to worry about permissions, they're a hassle and I don't think the risks are very high. If that doesn't work for you then you'd have to figure that part out.
2) Message boards : Questions and problems : Errors Moving Data Folder (Message 108192)
Posted 22 May 2022 by Nick Name
Post:
You should be able to copy everything over to whatever location you want, and then run the client in that location using boinccmd.

https://boinc.berkeley.edu/wiki/Boinccmd_tool

Here's an example of a batch file I use on a Windoze machine.

"c:\Program Files\BOINC\boinc.exe" --gui_rpc_port 31450 --dir O:\GPU\AMD\AMD31450
3) Message boards : GPUs : One Nvidia GPU unable to process after a couple of days. (Message 108191)
Posted 22 May 2022 by Nick Name
Post:
Older BOINC versions had problem with Einstein, under some circumstances they would download more work than could be completed by the deadline. This might be your problem, it explains why things work correctly for a couple days after you delete tasks and then the problem comes back. I'm certain the excessive work problem affects the 7.16 series. The easiest thing to do is update BOINC to a newer version and see if that fixes it.

Keep in mind some of those Einstein GPU tasks actually require more than one thread, so your CPU usage settings might not match the workload like you think.
4) Message boards : Questions and problems : Is there a way to get BOINC to give priority to Tasks with short "time to deadline"? (Message 107699)
Posted 3 Apr 2022 by Nick Name
Post:
I don't know if you have any control over Science United tasks, but you can force standard BOINC tasks to run by doing the following.

1) Set the project you want to to run now to NNT (No New Tasks).
2) Suspend the other projects or pause all other tasks. Click Show All Tasks in the manager to see every task in the queue.

This will start the project tasks you want. If you don't set that project to NNT you may find it downloads more work for that project. That's counterproductive so make sure to set to NNT first.

You should set a low cache to keep this from happening. If you're using local settings in the manager, go to Options -> Computing Preferences -> Other = Store At Least. Normally I don't like to use more than a day. Things should balance out over time, think weeks rather than days, but you can help things along by setting a minimal cache.

CPDN is an outlier, it's the only project with such extreme deadlines and in my opinion shouldn't be run with other projects for that reason. You will not have any control over that with SU, and I don't recommend running SU either.
5) Message boards : Questions and problems : Black screen / Win Update? BOINC not sharing nicely. (Message 103603)
Posted 19 Mar 2021 by Nick Name
Post:
In addition to Dave's comments, it's good practice to lower CPU load if you are running GPU work to make sure GPU apps also have a full thread available. I like to keep one available for OS overhead as well. Using 8 threads / one GPU as an example, set Percentage of CPUS to 75%.

You can add an exclusive app in BOINC and it should suspend work while that app is running. In the manager, go to Options -> Exclusive Applications. You can also set a time schedule to allow for Windows updating or other tasks like backups, that's under Options -> Computing Preferences.
6) Message boards : GPUs : How to increase GPU performance? (Message 103113)
Posted 21 Feb 2021 by Nick Name
Post:
Temperature wise is under a controlled condition, with fans set at a specific speed, and at a controlled ambient temperature.
Wattage is obvious.
Other indicators (software) often see only a part of the GPU. Eg: if your 64bit Double Precision processors are working at 100%, but none of the shaders are, the GPU may be showing 100% activity, however it may be pulling only 1/3rd of the amount of wattage, and as a result, runs very cool.
The GPU frequency may even be boosted to max, but if you're not using the 32bi / 16 bit shaders, then you're not using the GPU to it's fullest.

As far as temperature not being a good indicator, temperature is directly related to wattage consumption, so yes, in a controlled environment, temperature can very accurately tell you how much of the GPU is used!
Maybe not as a direct indicator, but it can show you as an average (as temps fluctuate slower than wattage ratings).

The best way to see GPU utilization is read it's wattage in Linux is easy (nvidia-smi).

Not everyone is running systems in climate controlled rooms. Years ago I had a dual axial fan 760 that ran quite hot, to the point it was severely thermal throttling. That GPU cooler was not a good choice for that system. If I were only monitoring temps I would have thought the card was working extremely hard, and been very confused when my output was not what what it should have been. Temperature is not a good substitute regardless of the factors you state, although it's important to monitor for system health.

I find it very odd that you trust nivdia-smi for temp and power reporting but not the actual utilization rate, unless your argument is that power consumption is equivalent to utilization . That's not true either as the card can run at or near 100% without maxing out its power draw. If the compute load is 100%, the card is working to its capacity regardless of the power rate. The application's demands are what matter for utilization reporting, not that the GPU is performing every type of calculation it's capable of.

My preferred GPU monitoring app for Nvidia on Linux is nvtop. It shows you everything discussed here and more, including VRAM usage, so everyone can see their preferred metrics. To be clear, I do monitor the power rate but that's mainly because I set power limits. If there's a driver crash the power limit has to be reapplied, at least on Windows.
7) Message boards : GPUs : How to increase GPU performance? (Message 103074)
Posted 18 Feb 2021 by Nick Name
Post:
I wouldn't recommend using those for compute purposes.
8) Message boards : GPUs : How to increase GPU performance? (Message 102991)
Posted 15 Feb 2021 by Nick Name
Post:
Windows task manager, GPU-Z, HW Monitor, MSI Afterburner, ....
All these programs can't really tell how busy your GPU is running at.
They will give you an estimate, but it's often dead wrong.
The closest thing you can measure to GPU utilization, is wattage or temperature.
If the GPU is 100% loaded, it'll draw x-amounts of watts, and with the fan at 100% will run at Y-degrees temperature.
In Linux, the wattage can be directly read via the nvidia-smi command.

Often dead wrong? There's no reason to think these programs aren't accurate as they're getting the data from the driver. nvidia-smi is part of the driver package, also available in Windows and all these values including utilization can be verified with it if one desires. I've done it, and GPU-Z and Afterburner are accurate, as well as System Information Viewer in my testing. Afterburner in particular is really just a nice GUI for nivida-smi; if we can trust it to set clock speed and power limits (we can) we can trust it to report utilization. If you're arguing the driver isn't reporting accurate data that's different, but let's see the proof. There are bugs from time to time, but that's expected with software.

Temperature is not a reliable indicator; ever seen the results from an incorrectly mounted or poorly functioning cooler? Conversely a well functioning system can keep things frosty, relatively speaking, and temps might be "low". This could mislead you into thinking the GPU isn't working that hard when it actually is, especially for lower powered GPUs like the OPs. I wouldn't strictly trust the wattage reported from software either, but this is just muddying the waters anyway as it relates to the original question.
9) Message boards : GPUs : How to increase GPU performance? (Message 102980)
Posted 12 Feb 2021 by Nick Name
Post:
You're probably not seeing the the true GPU utilization. I don't have Win10, but there's a setting in its task manager for seeing GPU compute loads. That's what you need to look for. You can also use a 3rd party app like Afterburner or my favorite, System Information Viewer. I'm sure your GPU load is actually much higher than 5%, but if it isn't then your CPU is probably over loaded. Reduce the number of CPU jobs running for better output.

You can't directly compare how projects run, as different apps can behave quite differently. It's true in all cases that running concurrent tasks will take longer than running one. The test is to see how it works over time. As a simple example, if a single task takes five minutes but running two at once takes less than ten minutes, that advantage adds up over time. In such a case you can keep scaling up until you hit the limit. You still have to make sure you have enough CPU support, Nvidia apps in particular tend to like having a full thread available.
10) Message boards : Questions and problems : Tasks freeze until reset (Message 102226)
Posted 18 Dec 2020 by Nick Name
Post:
Three possibilities come to mind:

The machines are overtaxed; there are too many BOINC jobs running, or too many background tasks running or some combination of both. Reduce the number of jobs until you find the limit.

Jobs are using up all available RAM and hitting the swap, slowing things way down. If the machines aren't locked up and are fairly responsive to user input you can probably rule this out.

The project app is buggy. It might be stuck and falsely reporting progress, check your CPU load to see if it's what you expect. If the tasks are hanging as you describe and everything else is working properly, that needs to reported to the project. Check the project forum to see if others are reporting this problem. It's also possible that this is a new app and BOINC's time estimate is way off. The solution in that case is to let it run and the time estimate should correct itself eventually.
11) Message boards : Questions and problems : suspended CPU usage (Message 102119)
Posted 13 Dec 2020 by Nick Name
Post:
BOINC doesn't dynamically manage resources in that way. <suspend_cpu_usage> means if the system load - not counting BOINC - is above that percentage, stop computing. There's no way to do what you want unless you can script something using boinccmd.

re. global_prefs, it will be in the BOINC data folder, however if you are changing settings in the manager the file you want to look at is global_prefs_override.xml. Parameters are set with the BOINC manager so there shouldn't be any need to mess with it. If you do decide to tinker with it outside of BOINC, use a plain text editor and make sure it's set as an XML file.
12) Message boards : GPUs : My ATI Radeon HD 5400 R5 210 series (Cedar) Will not compute with milkyway@home, astroids@home and Moo! Wrapper, what do i do. (Message 102114)
Posted 13 Dec 2020 by Nick Name
Post:
Asteroid's GPU work is Nvidia-only, as the app is written in CUDA. The project is down right now anyway.

MilkyWay requires double-precision, aka FP64 compute capability. I don't think your Radeon has it, and that's why you can't get work there.

I don't know what Moo is doing these days, but my guess again is that Radeon isn't compatible, maybe it's just too old.
13) Message boards : GPUs : Any projects use AMD Radeon Pro Vega II OpenCL: 1.2 (Message 100776)
Posted 17 Sep 2020 by Nick Name
Post:
Your GPU should work on MilkyWay and Einstein. It may also work on PrimeGrid, that project does support Mac CPU but I'm not certain if GPU on the Mac works there.
14) Message boards : GPUs : Boinc won't find GPU any more (Message 100766)
Posted 16 Sep 2020 by Nick Name
Post:
I would guess that Windows has updated something and broken the driver, regardless of your settings. I'd try the simple thing first, download and install your preferred driver from AMD.
15) Message boards : GPUs : Beside ''1'' in cc_config.xml what else could be the problem (Message 100765)
Posted 16 Sep 2020 by Nick Name
Post:
You need to use a / to complete the options section.

<cc_config>
<log_flags>
<task>1</task>
<file_xfer>1</file_xfer>
<sched_ops>1</sched_ops>
</log_flags>
<options>
<use_all_gpus>1</use_all_gpus>
</options>
</cc_config>
16) Message boards : Projects : nanoHUB_at_home has failed every single task -- work time exceeded (Message 99955)
Posted 14 Jul 2020 by Nick Name
Post:
First, I suggest you dump Science United and use the standard BOINC manager. That way you will have full control over what's running on your machines. You could then detach from the project permanently or set the project to No New Tasks until the problem is solved.

This sounds to me like the work is never actually running. If you are successfully running other VirtualBox work then this problem needs to be reported to the project. I'd at least check their forums for reports of problems / solutions.

If you aren't running other VB work, you should check the excellent LHC guide for VB work to make sure your machine is setup correctly. The info about specific VB versions is a bit dated but the information overall is good. I also like to enable the VB window when I have a problem, I can usually see what it is from the log that's shown there. To do that, go to your cc_config file and set the vbox_window to 1.

<vbox_window>1</vbox_window>
17) Message boards : GPUs : PC with two GPUs (Message 99669)
Posted 7 Jul 2020 by Nick Name
Post:
Rosetta work is probably taking priority due to shorter deadlines. Try reducing the number of CPU threads BOINC can use.

Options -> Computing Preferences -> Usage Limits (Use at Most % of the CPUs)

Lower this so that BOINC uses at least one less thread than it is now, if you don't get the desired result keep reducing it until you do. BOINC will over-commit the CPU to run GPU work. If you have eight threads available, it will run eight CPU tasks plus as many GPU tasks as it can. Einstein GPU work likes a full thread too, so by lowering the Use at Most % you are - counter intuitively - freeing up resources for GPU work.
18) Message boards : Questions and problems : Possible issue. (Message 99546)
Posted 29 Jun 2020 by Nick Name
Post:
I'm with the OP on this. The point of selecting subprojects is that you only want to run those, not any others, and if that's the only project you're running it should be obvious that you're ok with an idle machine. How many times in these and project forums have we seen people asking why they're getting work that they didn't select? Most of time it's because they overlooked the "accept other work" checkbox, which is easy to miss if you're not aware of it. Frankly I'd be fine with the "accept other work" going away completely. If you're worried about idle machines the sensible thing to do is select everything.
19) Message boards : Questions and problems : BOINC always "Suspended - computer is in use", when not in use (Message 98570)
Posted 17 May 2020 by Nick Name
Post:
There have been a couple reports on the F@H forums about a similar problem with that client not detecting when the system is idle. Multiple distros are affected so I suspect there's been a change in newer kernels that causes this in some cases. I have no idea where to look or how to fix it though.
20) Message boards : Questions and problems : CPU Temps (Message 98319)
Posted 4 May 2020 by Nick Name
Post:
In the meantime, if your CPU fan isn't running fast enough you might look to change that in the BIOS. Most of the boards I've seen have the fan speeds optimized for lower noise, not performance.


Next 20

Copyright © 2024 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.