2019 Formula Boinc Sprints

[crashtech]

I'd be curious to try the avx app on my Ryzen 3700X CPUs. Right now they are running the sse2 app.

 

[StefanR5R]

In the past, TN-Grid's various application versions did not perform differently. Any differences in GFLOPS that were seen at the "application details" page of a host were down to random variations. I looked again now, and this is still the case with the current application versions (sse2, avx, fma) on my Broadwell EPs.

(I also have one Haswell on it, but its GFLOPS per application version are so wildly disparate that I dismiss these values entirely. The good thing about the BDW-EPs is that I have a good number of "hosts" running, such that it is clear from comparison between them that differences in GFLOPS of the application versions are just noise.)

Also, the BDW-EPs run these applications at their non-AVX turbo clock, not at the lower AVX turbo boost 2.0 clock. That is, vector units usage of these applications, including the fma one, is low enough for the processor firmware to decide to keep going without AVX(2) clock offset.

 

[StefanR5R]

We stepped up one rank in the TN-Grid marathon. :-)

And we are tied right now with SG overall.

 

[biodoc]

I'd be curious to try the avx app on my Ryzen 3700X CPUs. Right now they are running the sse2 app.

Like @StefanR5R on his broadwells and Haswells, my 3700X does not show significant differences between sse2, avx, and fma apps.

gene@home PC-IM 1.10 x86_64-pc-linux-gnu (avx)
Number of tasks completed    57
Max tasks per day    157
Number of tasks today    0
Consecutive valid tasks    57
Average processing rate    7.07 GFLOPS
Average turnaround time    0.37 days
gene@home PC-IM 1.10 x86_64-pc-linux-gnu (sse2)
Number of tasks completed    72
Max tasks per day    172
Number of tasks today    5
Consecutive valid tasks    72
Average processing rate    7.07 GFLOPS
Average turnaround time    0.55 days
gene@home PC-IM 1.10 x86_64-pc-linux-gnu (fma)
Number of tasks completed    73
Max tasks per day    173
Number of tasks today    25
Consecutive valid tasks    73
Average processing rate    7.19 GFLOPS
Average turnaround time    0.44 days

 

[StefanR5R]

TSBT have probably completed their work in league 2 already, and switched to one or another league 1 team.

Edit, maybe they can still catch up with Gridcoin.

Turns out it's one and then another team, not one or another.

Just 5 more hours to go, but maybe they can still do it.

We stepped up one rank in the TN-Grid marathon. :-)

And we are tied right now with SG overall.

...and this is already yesterday's news now. :-)

 

[biodoc]

RKN and LAF passed SG in the sprint so now we are 4 points ahead of SG overall. I checked the LAF forum and at least a couple of members are participating in the sprint. This year will be the worst finish by far for LAF in FB. Maybe that's why Sebastien has lost interest in running FB.

 

[Orange Kid]

TN-Grid final
Fourth place it is. We also moved up two spots in the marathon gaining 3 points and passed SG in the overall standings!!!



From FreeDC



The Singapore F1 results

• Sebastian Vettel 1:58:33.667
• Charles Leclerc +2.641
• Max Verstappen +3.821

 

[StefanR5R]

Thanks for the stats!

@Nightmare2, @VietOZ, good job taking first in this one (and 3rd as well).

Overall points of the TN-Grid project according to boincstats:

2019-09-22 -- 6.2 M
2019-09-21 -- 4.9 M
2019-09-20 -- 7.5 M
during the two weeks before -- 2.7...3.9 M​

I take from this that TN-Grid's work generator can create tasks for no more than 5...6 M PPD these days (was 6...7 M PPD last year). On September 20, work from more than one day was validated.

 

[crashtech]

It's LHC!

Formula Boinc

Formula Boinc Championship results

formula-boinc.org

 

[StefanR5R]

I come to the realization that LHC is actually more broken than I remember.
Example: https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4686
(limit on tasks in progress == number of configured cores per task)

What a garbage pit.

 

[crashtech]

I don't know how to make LHC work right. It never fully utilizes the cores, gives me various error message, and/or waits for memory when plenty is available.

 

[StefanR5R]

It never fully utilizes the cores,

I agree. I suspect this could be mostly fixed by running Atlas-Native with a low thread count per task. But I am not going to install and operate LHC's distributed filesystem (...natively, that is. The VM jobs are most likely running that same filesystem too.), hence I will never find out for myself.

gives me various error message,

Yes. Just now I had all vbox based jobs go into "unmanageable" state on each of my computers. This incidence was likely caused by me downloading and/ or installing system updates, which I haven't done on these OSs for a while. Maybe it was caused by how the package manager installed the updates, or it was caused by the high internet bandwidth use during the downloading.

On the first computer, I restarted the boinc client, and this resulted in the previously postponed vbox jobs to fail with computation error. I rebooted the other computers, and on these the jobs resumed were they stopped and appear to continue to run properly.

and/or waits for memory when plenty is available.

Did you remember to set "When computer is in use | is not in use, use at most [ . . . ] %" to something like 90 %?
Still, it seems to me that the vbox jobs allocate fewer memory than what the boinc client is told that their (peak) RAM usage will be. This surprises me; I thought vbox allocates the whole RAM, which a VM is meant to get, right away at the start of the VM.
For me that means 90 % RAM granted to boinc = 60 % actually used. (Running the native application with its lower RAM requirement would solve this AFAIK. But again: No, thank you very much, I'm not going to manage a distributed filesystem on my computers.) Two boinc clients could be used to work around that, but then extra caution needs to be applied to not to exceed the available RAM.

LHC@home is certainly the worst integrated one of the wrapper projects in the boinc world.

 

[Ken g6]

Now I can't get BOINC to detect my virtualbox install. There's no line detecting it in the BOINC log. I know it's installed - I'm running a separate VM right now.

 

[StefanR5R]

Is user "boinc" (or whichever user runs the client) member of group "vboxusers"?


----
Linux users who for some reason get nostalgic and long for a Windows 3.1 feeling of fragile systems which crash when you look at them wrong, simply need to use VirtualBox.


----
Credit relative to CPU time × measured application GFLOPS (compared between hosts with identical hardware and application config and very similar CPU times and app GFLOPS), is about as random as the scattering of particles when Large Hadrons Collide. This is YAFU level, if not worse.

 

[Ken g6]

Is user "boinc" (or whichever user runs the client) member of group "vboxusers"?

I think so. That's the first thing I did. `sudo usermod -a -G vboxusers boinc`

 

[StefanR5R]

Are the kernel drivers loaded? On the OS which I am currently using, Opensuse, this is done via vboxdrv.service. There is also a vboxautostart.service. Both are currently "loaded active exited".

 

[Orange Kid]

LHC@home stats day one+
In third at this writing, have seen us in second.




From FreeDC

 

[Orange Kid]

LHC@Home day two
Second place and looking good. Going to be a race to the finish as bunkers start to drop. Hang on to your hats.




From FreeDC

 

[Orange Kid]

LHC@Home final

Second Place!!!! Nice sprint and congrats to all.



From FreeDC



The Russia F1 results

• Lewis Hamilton 1:33:38.992
• Valtteri Bottas +3.829
• Charles Leclerc +5.212

 

[motqalden]

Nice work and thanks for the friendly boost!

 

[Orange Kid]

Universe@Home is up next.

 

[Orange Kid]

Hold your water, things don't look quite right at FB.

EDIT: the universe seems to be down at the moment.

 

[crashtech]

Hmm, there should be scores up by now, right!?

 

[biodoc]

Yes, I'm pretty sure that means FB doesn't have baseline (starting) stats yet so hold your bunkers until you see a list of teams with 0's.

 

[Orange Kid]

Universe is back but FB is still FUBAR.