Posts by ProDigit

1) Message boards : The Lounge : No more pandemic unemployment boost, can't afford Boinc anymore. (Message 100210)
Posted 5 days ago by ProDigit
Post:
Sad and the fact you have much company is no solace. I could get political but that might get me a lifetime ban.

I think everyone would agree, that at least something, you know?
Just until there are jobs.
But no one ever said I could stop paying taxes, because I've been working for so long without ever needing govt assistance :/

Another friend said, they get more in holiday pay, than what most furloughed and laid off Americans get as unemployment for those months..
2) Message boards : The Lounge : No more pandemic unemployment boost, can't afford Boinc anymore. (Message 100208)
Posted 6 days ago by ProDigit
Post:
I was recently furloughed for an undisclosed amount of time, after nearly 9 years of working in the same company.
Now that the $600/Mo Covid package is no longer there, I'm going to have to totally unplug Boinc.
I have to provide for my family, you know...
And with no end in sight, and my company may or may not open doors anymore...

Today was the last day.
3) Message boards : Projects : Is this an ok way to benchmark Azure VMs? (Message 100188)
Posted 7 days ago by ProDigit
Post:
On all systems, MS Azure, Google Cloud Computing, and Amazon AWS, a system gives less score than 1/3rd of the actual hardware.
I don't know why this is the case, other than them sometimes cutting performance to the CPU cores.
I removed all my cloud accounts, bought a few Atomic Pi units, and though they run at only 1,69Ghz, they still outperform the Intel Xeon 2,1Ghz quadcores by a margin of almost 3x.

What's worse, the shared cores are being cut sometimes for weeks on end running at below 100%.
Sometimes 80%, sometimes 50%, sometimes they run for weeks/months at 1% if unattended.
4) Message boards : Questions and problems : boinc not suspending GPU while in use (Message 100162)
Posted 9 days ago by ProDigit
Post:
Fedora, though being well aged for a Linux distribution, is not one of the most common operating systems either.
It's based on Redhat, which also CentOS is. I wonder if those 2 are also affected by it?

I've never really played around with RPM based linux versions, but most of the Debian and Ubuntu based operating systems seem to work quite well.
5) Message boards : News : New BOINC project studies machine learning (Message 100086)
Posted 18 days ago by ProDigit
Post:
System requirements (64 bit CPU only):
Not supported:
GPU

Too bad, this kind of project could gain so much traction if GPUs were used...


If it is similar to CPDN where each calculation depends on the previous one then it is possible GPU won't give much of a boost if any. Of course it may well be able to make use of parallel processing. I am just putting forward a possible reason.


Only if the project depends on Double Precision.
Then an RTX2080 @ 300W is about 10x faster than a 15W dual core Celeron CPU at the same frequency.
GPUs can use VRAM to store the result, which they re-read, and re-process through the shaders/cores, without needing to pass through the PCIE bus again.

If a project can run on Single Precision (or 32bit FPP), an RTX 2080 Ti is about 1.000-1.500x faster (in terms of Flops) than a dual core celeron running at the same frequency.

Long WUs may require a lot of VRAM, especially if tailored to the GPU they're feeding.
An RTX 2080 Ti has ~4350 cores and 11GB of VRAM, that means each WU gets less than 2,5MB of VRAM, or there must be some parallel optimizations allowing blocks of cores (shaders) to run simultaneously sharing VRAM memory blocks.

If Half precision (16bit) worked, Nvidia and AMD are working on doubling their half precision process rates, by processing 2 of them in their cores. Rather than running 1 instruction per core, they can fit 2 of them through an algorithm.
Not sure if this process will be implemented in current line of GPUs.
But to give you an idea, current AI runs at Quarter (8 bit) or half(16bit) precision, and uses between 10 to 300 cores.
Can you imagine the processing power if ran through an RTX 2080 Ti (with over 4000 cores)?
Should boost learning algorithms and AI to near real time.
6) Message boards : Questions and problems : Laptop stops crunching WUs when lid closes (Message 100085)
Posted 18 days ago by ProDigit
Post:
You may have answered your own question - some laptops run very much hotter with the lid down than the lid up, and thus, despite all our attempts to keep them running the thermal management kicks in, over-rules our settings and depending on the laptop either seriously reduces performance or stops working. One thing to remember is that laptops have highly optimised thermal management systems, and the vast majority of these systems rely on the exposed surface of the keyboard to shed some of the heat, and putting the lid down puts what is effectively an insulating layer into the cooling path.

No in the past it would crunch through data at full speed, as the fan cools down the CPU enough, that closing it wouldn't matter.
In the past I did only GPU crunching, now I add CPU to it.

With GPU crunching, adding 1 CPU thread = 60c, plenty cool even with lid closed 60% CPU utilization @~2,2Ghz.
GPU + 2 CPU threads = 75C (just a little thermal throttling going on ~1,8-1,9Ghz down from 2,59 max with CPU only and without GPU)
2 CPU threads + GPU also show 100% CPU utilization in Taskmanager, so I don't add more than that.

75C with lid closed should still not thermal throttle.
Windows may however, lower C-state with the lid on, where as before it would just lower C-state when the laptop went to sleep.
(sleep/hibernation/log off is disabled on my laptop).
7) Message boards : Questions and problems : BOINC Not Suspending during Computer Use (Message 100084)
Posted 18 days ago by ProDigit
Post:
Yes, Boinc needs a certain CPU percentage of non-boinc activity, for a certain amount of time, before the limiter kicks in on the CPU.
GPUs (I think) will just continue to crunch full blast anyway.
8) Message boards : Questions and problems : Missing start tag in cc_config.xml (Message 100083)
Posted 18 days ago by ProDigit
Post:
Ok, the issue disappeared by deleting cc_config.xml.

It's really odd, because I haven't touched that file, when the error happened.
It happened after adding MLC@Home, and modifying it's app_config.xml file.

I haven't adjusted app_config.xml, just deleted cc_config.xml, since I don't have anything useful in there anyway, and that did the trick.
9) Message boards : Questions and problems : Missing start tag in cc_config.xml (Message 100054)
Posted 18 days ago by ProDigit
Post:
It appears to affect only the Windows version.
Linux seems to work fine with it.

I've used Notepad++ to see if any strange characters were present, but nothing so far.
Besides, it would be odd for Boinc to refer to cc_config, when there's an error in app_config.
The default cc_config was blank, I might revert it to a blank file again; and for the mean time, adjusted my CPU settings to 50 or 75%, as MLC is the only project currently running (next to collatz).
10) Message boards : Questions and problems : Missing start tag in cc_config.xml (Message 100045)
Posted 19 days ago by ProDigit
Post:
I used nano. It usually works fine.

app_config.xml content:

<app_config>
<app>
<name>mlds</name>
<max_concurrent>1</max_concurrent>
</app>
</app_config>
11) Message boards : Questions and problems : hours limitations on gpu use (Message 100043)
Posted 19 days ago by ProDigit
Post:
I don't know of any easy way to do it.
What you can do, is limit the power usage of the GPU with programs like MSI Afterburner, EVGA XOC or Precision X(OC), or Zotac Firestorm; or do "sudo nvidia-smi -pl xxx' in Linux ('xxx' being the desired power value, eg: 145 for 145W); and lower the TDP by about 20 or 30%.

That way your GPUs will run a lot cooler all over.

For any script to automatically do this, is a bit difficult.
The controls on Windows are via GUI, and aren't easily adjusted with a macro, but it could be done.
12) Message boards : Questions and problems : Laptop stops crunching WUs when lid closes (Message 100042)
Posted 19 days ago by ProDigit
Post:
odd.
I can't think of anything else that happened to it.

I can tell that it stops working (or at least, stops working at full speed), when I run HWmonitor.
HWMonitor registers 75C under full load (I have a tiny USB fan mounted on the back, blowing in air in the chassis, as this is a fanless system).
When I open up the laptop again, it registers 35-40C.
It does appear like some WUs have been processed, however, the laptop feels cool, not warm like usual.

Nothing else has happened to it.

I just now (after the incident), updated Boinc to 7.16.7; hoping it'd fix the issue.
Will see tonight, as I'm needing it now for work.
13) Message boards : Questions and problems : Missing start tag in cc_config.xml (Message 100041)
Posted 19 days ago by ProDigit
Post:
I'm testing out Boinc 7.16.7 (x64) on Windows 10 home (64 bit).
It all works fine, however, when I tried limiting a project to 2 simultaneous WUs in app_config, Boinc doesn't honor the values.
In the log there's a note saying 'Missing start tag in cc_config.xml'.
Thinking that cc_config.xml needs to be set before app_config.xml can be set, I populated it with this:
<cc_config>
<options>
</options>
</cc_config>


As I'm not needing any special configurations there.
But I still get the error.
Instead of telling me 'name is wrong in app_config.xml' or so, it tells me there's something wrong with cc_config.xml?
14) Message boards : News : New BOINC project studies machine learning (Message 100020)
Posted 20 days ago by ProDigit
Post:
System requirements (64 bit CPU only):
Linux
Windows

Future support:
OSX
Linux / arm
Linux / ppc64le

Not supported:
GPU

Too bad, this kind of project could gain so much traction if GPUs were used...
15) Message boards : Questions and problems : Laptop stops crunching WUs when lid closes (Message 100019)
Posted 20 days ago by ProDigit
Post:
Same laptop that has been running WUs for months on end, now with a Windows update (Windows 10 Home 64-bit), no longer crunches, when the lid closes.

I went through all the power settings, but couldn't find any setting that was wrong.
Even set back the 'lid close - do nothing' to 'sleep' or 'shutdown', and back to 'do nothing', but to no avail.

Was wondering if someone else has experienced the same thing?
16) Message boards : Questions and problems : LHC@Home on BOINC Android Could not connect (Message 99991)
Posted 22 days ago by ProDigit
Post:
https://photos.app.goo.gl/hawXcG8Fh3SCCkWE9

I am running BOINC 7.4.53 on Android 9

When I try to connect to LHC@Home I get the error Could not connect.


LHC has no WUs for android as far as I know.
I've never had any on my Android 6, 7, 9, and 10 devices.
Best is to choose another project for your device.
Also, don't run Boinc on a phone you still need. The load will definitely destroy the battery over time.
I myself run them on TV boxes (no battery), and 2 old cellphones that already have a blown battery (screen pops out).
17) Message boards : Projects : Folding@Home and BOINC (Message 99961)
Posted 26 days ago by ProDigit
Post:
In short, Fah couldn't run in Boinc; since FAH depends on quick return of WUs (usually the deadline is below 2 days), and sharing resources would end up in issues.
???
I run a project whose deadline is 24 hours. Works fine.

I mean, FAH has a Quick Return Bonus, which is an additional PPD rating on top of the regular rating that Boinc doesn't have.

FAH also can't just assign any WU to any GPU, like in Boinc.
They have specific WUs that are made for a specific range of GPUs, in a way to make maximum use of the GPU's hardware.
This is one thing that Boinc doesn't yet have.
It's not a thing for BOINC to have. BOINC doesn't use the CPU or GPU, that's done by the projects and if a project would dictate it can only run on certain GPUs then only those GPUs can be used. See GPUGRID. See Milkyway@Home. See Minecraft@Home.

True, but these projects make use of available hardware that can be shared with other projects.
FAH is more trying to completely and unanimously have control of the hardware given.
Eg: if you have 2 GPUs assigned to FAH, FAH won't share this with other projects.
It's not made for WUs to be paused in favor of other WUs.
WUs also don't stay on the host. As soon as a WU is nearing completion, another WU is downloaded; and then seamlessly processed in a serial-kind of way.
With Boinc, one can stack upto (I believe) a good 16-20 days of WUs.

Each has it's benefits. For a wonky internet connection, BOINC is definitely preferred.
FAH puts a lot of emphasis on fast and modern hardware, a good (always on) internet.
While people with slow or intermittent internet can still donate, they usually suffer a large PPD penalty; which on boinc they don't.
In many cases it doesn't matter if a WU on Boinc is returned within the hour, or takes a few days to complete.
The PPDs are mostly the same for both (unless the project assigns additional PPD for the first one to complete the WU, and anyone following is just confirming or verifying the WU).
18) Message boards : Projects : nanoHUB_at_home has failed every single task -- work time exceeded (Message 99960)
Posted 26 days ago by ProDigit
Post:
Large RAM WUs usually are via docker or VM.
If your OS is Windows, it could have to emulate Linux through an emulation layer.
If you don't have enough RAM, and the system is reading from SWAP, it may cause the WU to process very slowly.
19) Message boards : Projects : Folding@Home and BOINC (Message 99947)
Posted 26 days ago by ProDigit
Post:
Fah has strucutally different methods from Boinc.
Not limited to projects assignments, and PPD counting system (they use QRB PPD on top of the standard PPD).
In short, Fah couldn't run in Boinc; since FAH depends on quick return of WUs (usually the deadline is below 2 days), and sharing resources would end up in issues.
FAH also can't just assign any WU to any GPU, like in Boinc.
They have specific WUs that are made for a specific range of GPUs, in a way to make maximum use of the GPU's hardware.
This is one thing that Boinc doesn't yet have.
I can run a GT 1030 and an RTX 2080 Ti, and the configuration to get the right WUs to work on the right GPUs would take too much user interaction, that a normal person wouldn't be able to figure out.
FAH is aimed towards someone who can run windows or linux, install FAH, and run a program.
Not fiddle with programs until they work optimally.
20) Message boards : Questions and problems : Other Volunteer Computing Ap error (Message 99946)
Posted 26 days ago by ProDigit
Post:
It sometimes shows, when the Boinc app is being loaded twice at the same time.
All you need to do, is close the app, and reopen it.


Next 20

Copyright © 2020 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.