Am I missing something, or is there a bigger picture here...

GWGeorge007
GWGeorge007
Joined: 8 Jan 18
Posts: 352
Credit: 512,651,691
RAC: 2,860,812
Topic 224649

I've been following the posts in Cruncher's Corner for E@H and though I've been silent (until now) does anybody else, besides me, see a bigger picture than what is going on right now with E@H?  What I mean is I've also been following the cruncher's posts for MW@H and U@H, and what I'm hearing there is also some sort of discrepancy, at least, within their projects also.  I do not know what effect, if any, there is with the multitude of other projects in BOINC as I don't frequent the forums as much as others do, or that I would like to.

Is it possible that the underlying problem (which has yet to be fully resolved) is actually something a little bigger, like in the BOINC application 7.14.2 being used now, and possibly even an older version?  I feel like solving the E@H problem(s), or any problem within the other projects, is like putting a bandaid on a broken arm.

I'm just tossing this up as food for thought.

But it certainly has me thinking... does it make you think too?

George

Ian&Steve C.
Ian&Steve C.
Joined: 19 Jan 20
Posts: 853
Credit: 4,614,791,868
RAC: 31,512,635

it might help if you were

it might help if you were more specific in what you're referring to. what discrepancy(ies)?

_____________________________________________

San-Fernando-Valley
San-Fernando-Valley
Joined: 16 Mar 16
Posts: 198
Credit: 4,391,897,402
RAC: 6,040

... would definitely help -

... would definitely help - what discrepancies?

GWGeorge007
GWGeorge007
Joined: 8 Jan 18
Posts: 352
Credit: 512,651,691
RAC: 2,860,812

Maybe discrepancy was a wrong

Maybe discrepancy was a wrong choice of words.  You can insert your own choice of words, but here is what I found that made me think.

In Milkyway there is a thread regarding “Updated GPU Requirements” located here:

https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=3188

Ricarter96 mentions in his post from Jan 25, 2021 (message 70434): “Have been running MW on an Asus mobo, Intel Core i7-3930K Hexa-Core CPU, Windows 10 Enterprise x64, NVIDIA GTX 1070 Ti, and its Cuda cores were crunching numbers very well. Last week I built a new system especially for targeting this kind of task and others like it. Moved my NVIDIA GPU to the new system, installed MW but now, no Cuda number crunching. [snip]”

Then Ricarter96 mentions in his post from Jan 27, 2021 (message 70459): “The Advanced dialog clearly states that GPU crunching won't work if the app is installed as a service. [snip]” 

This is the last post in the thread as of the time that I write this which leads me to think that he is correct as it has not been corrected or challenged.

 

In Universe there is a thread specifically pertaining to running Universe in Hyper-V.  It would be in your best interest to read the thread, including links, either before or after ‘glancing’ through it for relevance.

https://universeathome.pl/universe/forum_thread.php?id=551

I realize that running Universe in Hyper-V does not appear to be relevant to the E@H problems at hand, but maybe (just maybe?) it does play a role in it (them?).  What concerns me is the mention of the updates to Windows 10 Pro and Ubuntu 20.04 indicating that they may (also) have a role to play in this too.

 

Now I may be all wrong in posing this question to this forum (maybe I should just shut-up and be quiet), but it seems like there maybe some underlying link to what’s happening in E@H that hasn’t yet been diagnosed or thought of yet.  Maybe I put too many “maybe” in this post too.

I only bring this up as food for thought that there MAYBE an underlying link to the BOINC application 7.16.11 being used now for Windows 10 64bit and 7.14.2 for 32bit, and also 7.4.22 for Linux, both 64bit and 32 bit versions.

Once again, I'm just tossing this up as food for thought.

George

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 1,304
Credit: 2,147,966,379
RAC: 11,749,433

The BOINC client HAS NOTHING

The BOINC client HAS NOTHING to do with running tasks. The science application is what runs tasks.

So, it matters not one whit which version of BOINC you are running.

The last issue with E@H was simply a problem with the science application for GR that uncovered a problem with the original Volta/Turing/Ampere application that necessitated a further modification in the TV application to handle the latest GR work units.

The only problem you can have with the BOINC client is whether it communicates with projects as you expect.

The Universe tasks run best on a Linux OS with the latest glibc libraries.  Windows does not use those libraries so is not as optimized.  You can get around the Windows liability by running Linux in a VM and you can achieve 95-100% of the same performance you get with native Linux.

Again, nothing to do with the BOINC client, rather with the science application. Please remember the science application for each project is what actually runs the work.  The BOINC client does nothing but communicate with projects to retrieve and send back work.

 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.