Warning - BigAdv: Check first 3 frame times.

Q

Qinsp

2[H]4U
Joined
Jan 7, 2011
Messages
2,154
I hit a new record.

Dual Xeon hexcore @ 3.0ghz = P8101 27:00 TPF.

It grabbed a P8101 and started folding at >40:00 TPF. ie - FAIL.

You just stop and restart, and all is well.

Does anyone have an idea what causes this?

It happens on HDDs/SSD's, 10.04/12.04, Intel and AMD, Dual CPU and Quad CPU, 6.3.4 and 7.3.x. So while it's random, it happens to all platforms I have. The CPU's all report at 100% usage, and the heat is up, so it's folding at full blast, just isn't working correctly.

But as far as I know, only P8101.
 
8101s are Stanford's little bundles of joy. Most 2P system can not fold them in time for bonus points. Until 8103 come out of beta, bigadv is not really doable on a dual hex-core setup..... unless maybe a SR-2 with things overclocked.
 
Qinsp said:
I hit a new record.

Dual Xeon hexcore @ 3.0ghz = P8101 27:00 TPF.

It grabbed a P8101 and started folding at >40:00 TPF. ie - FAIL.

You just stop and restart, and all is well.

Does anyone have an idea what causes this?

It happens on HDDs/SSD's, 10.04/12.04, Intel and AMD, Dual CPU and Quad CPU, 6.3.4 and 7.3.x. So while it's random, it happens to all platforms I have. The CPU's all report at 100% usage, and the heat is up, so it's folding at full blast, just isn't working correctly.

But as far as I know, only P8101.
Click to expand...

It can happen with any WU, but is much more problematic on large ones. It has to do with the way NUMA is handled in Linux. There is currently an active effort to recode the entire system, but the solution seems to be a ways off still.

Scheduling on NUMA systems, where it is highly important to keep processes and their memory together on the same node, is a longstanding problem area for Linux. In the 3.8 development cycle we got some necessary infrastructure to let developers start experimenting with solutions, but actually getting a handle on this problem will take a while yet.
Click to expand...
 
Whenever such issue is manifesting on 6.34 machines, please run fahdiag, pipe it into
pastebinit and paste resulting URL:

fahdiag setup:
Code: 
cd $HOME
wget http://darkswarm.org/horde/fahdiag/fahdiag.sh
chmod +x fahdiag.sh

running fahdiag:
Code: 
./fahdiag.sh | pastebinit
 
Was about to say :D .... SR-2 with dual Hex @ 3 or higher hits 8101 no issues...... As for oP - I noticed that, maybe thats why my WU EUE'd yesterday, not unstable OC (Its an SR2 so you never know ...)


402blownstroker said:
8101s are Stanford's little bundles of joy. Most 2P system can not fold them in time for bonus points. Until 8103 come out of beta, bigadv is not really doable on a dual hex-core setup..... unless maybe a SR-2 with things overclocked.
Click to expand...
 
I haven't seen it on my machine, dual hex xeons @3.2 (Stock), oh and it is good for all Bigadv projects - at the moment anyhow.
 
tear said:
Whenever such issue is manifesting on 6.34 machines, please run fahdiag, pipe it into
pastebinit and paste resulting URL:

fahdiag setup:
Code: 
cd $HOME
wget http://darkswarm.org/horde/fahdiag/fahdiag.sh
chmod +x fahdiag.sh

running fahdiag:
Code: 
./fahdiag.sh | pastebinit
Click to expand...


While the job is in Linda Lovelace mode, right? (choking on the big one).
 
My bad. :D

She got married and was Born Again? IIRC.

Dunno about marrying a gal who had more wieners in her than an Oscar Meyer warehouse.
 
402blownstroker said:
8101s are Stanford's little bundles of joy. Most 2P system can not fold them in time for bonus points. Until 8103 come out of beta, bigadv is not really doable on a dual hex-core setup..... unless maybe a SR-2 with things overclocked.
Click to expand...

I'll be a bit off topic here but 8103 aren't in beta. I've been picking them up now & then over the last couple of weeks and I don't run the beta flag.
 
Yes, the P8103's are in general release, I've folded a bunch of them. I'm crunching about 4 bigadv per day.
 
You must log in or register to reply here.
Back
Top