devlin85
27 hours? yikes! was it still just 3k credit 500,000 gflops project or was this some massive wu due to the large amount of cores?

Looks like the timing was way off, took less than 10 Hours to complete.
13321697  179948  5 Nov 2014 |  0:54:19 UTC 5 Nov 2014 | 7:46:52 UTC Completed and validated 24,644.46 491,876.70 13,040.22 Test application for CPU MD v9.02 (mtavx)
 

Name 24162-MJHARVEY_CPUDHFR-0-1-RND7385_0 
Workunit 10252539 
Created 3 Nov 2014 | 23:07:19 UTC 
Sent 5 Nov 2014 | 0:54:19 UTC 
Received 5 Nov 2014 | 7:46:52 UTC 
Server state Over 
Outcome Success 
Client state Done 
Exit status 0 (0x0) 
Computer ID 179948 
Report deadline 14 Jan 2015 | 0:54:19 UTC 
Run time 24,644.46 
CPU time 491,876.70 
Validate state Valid 
Credit 13,040.22 
Application version Test application for CPU MD v9.02 (mtavx) 

 
 

Stderr output
<core_client_version>7.2.42</core_client_version>
<![CDATA[
<stderr_txt>
BOINC wrapper for GROMACS.
Arg 0 [../../projects/www.gpugrid.net/mdrun_mtavx.901]
Arg 1 [--nthreads]
Arg 2 [20]
BOINC running with [20] threads
BOINC resolving [traj.xtc] to [traj.xtc]
BOINC resolving [topol.tpr] to [topol.tpr]
BOINC resolving [progress.log] to [progress.log]
BOINC resolving [state.cpt] to [state.cpt]
GROMACS: mdrun_mtavx.901, VERSION 5.0.2

GROMACS is written by:
Emile Apol Rossen Apostolov Herman J.C. Berendsen Par Bjelkmar 
Aldert van Buuren Rudi van Drunen Anton Feenstra Sebastian Fritsch 
Gerrit Groenhof Christoph Junghans Peter Kasson Carsten Kutzner 
Per Larsson Justin A. Lemkul Magnus Lundborg Pieter Meulenhoff 
Erik Marklund Teemu Murtola Szilard Pall Sander Pronk 
Roland Schulz Alexey Shvetsov Michael Shirts Alfons Sijbers 
Peter Tieleman Christian Wennberg Maarten Wolf 
and the project leaders:
Mark Abraham, Berk Hess, Erik Lindahl, and David van der Spoel

Copyright (c) 1991-2000, University of Groningen, The Netherlands.
Copyright (c) 2001-2014, The GROMACS development team at
Uppsala University, Stockholm University and
the Royal Institute of Technology, Sweden.
check out http://www.gromacs.org for more information.

GROMACS is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License
as published by the Free Software Foundation; either version 2.1
of the License, or (at your option) any later version.

GROMACS: mdrun_mtavx.901, VERSION 5.0.2
Executable: /var/lib/boinc-client/slots/1/../../projects/www.gpugrid.net/mdrun_mtavx.901
Library dir: /home/mjharvey/gromacs-gpugrid/linux64/share/gromacs/top
Command line:
  mdrun_mtavx.901 -ntomp 20 -nt 20 -x traj.xtc -s topol.tpr -g progress.log -cpi state.cpt

Reading file topol.tpr, VERSION 4.6.1 (single precision)
Note: file tpx version 83, software tpx version 100
Changing nstlist from 10 to 20, rlist from 0.9 to 0.928

Using 1 MPI thread
Using 20 OpenMP threads 
Compiled SIMD instructions: AVX_256 (Gromacs could use AVX_128_FMA on this machine, which is better)

The current CPU can measure timings more accurately than the code in
mdrun_mtavx.901 was configured to use. This might affect your simulation
speed as accurate timings are needed for load-balancing.
Please consider rebuilding mdrun_mtavx.901 with the GMX_USE_RDTSCP=OFF CMake option.


NOTE: The number of threads is not equal to the number of (logical) cores
      and the -pin option is set to auto: will not pin thread to cores.
      This can lead to significant performance degradation.
      Consider using -pin on (and -pinoffset in case you run multiple jobs).

starting mdrun 'Protein in water'
2500000 steps, 5000.0 ps.

Writing final coordinates.

               Core t (s) Wall t (s) (%)
       Time: 491825.292 24642.013 1995.9
                         6h50:42
                 (ns/day) (hour/ns)
Performance: 17.531 1.369

Thanx for Using GROMACS - Have a Nice Day

00:48:56 (11892): called boinc_finish

</stderr_txt>
]]>

Every now and then GPU Grid comes up with a CPU application. Every time they do, I give it a try and from a points versus time run they always come up short compared to other projects in the Bio/Medical category. While point return isn't a big part of why I will crunch a particular project it is part of the equation.

yodap

bill1024
I have turned in tens of thousands of WUs for WCG and not received a whole lot of points in return.
It's ok though, they have made some discoveries in curing childhood Cancer and fighting Aids.
If they can cure even just one kids Cancer, they can have all my points back.

Don't forget that Boinc scalps WCG points severely. About 1/7 of what WCG credits you . Still don't know why they do this!

 

The answer is quite simple and has been explained multiple times at WCG. WCG still accounts for points that pre-date their usage of BOINC. Specifically, when they were using UD client. However, instead of doing the smart thing and just starting everyone back to zero with a new client and vision, they decided to "eventually get around to fixing the issue". They have not done this. So, 7 WCG points is the equivalent to 1 BOINC point. This is so that those whom still have UD points have a consistent comparison of the work they do today with the work they already did with the United Devices work. For those whom just focus on WCG, it is not a problem. But for those whom compare between projects find the difference in numbers confusing and unneeded.

yodap
I use 326.80 on the 660ti. Good for both projects. In Linux, 319.xx is as high as I can get. Also works good for both.

I should probably post this in another thread, but as we prepare for the Cruncathlon.
Great performace on the 780 yodap:
 

devlin85
Seriously, hey yodap! What's your settings on that 780? I've gotten close to some of those times, but damn! Little help 

Umm I don't know. That was run yesterday and as you can see in my pic I borked 2 wu just before that while trying to update Windows 10. I was just giving the card a practice run for tomorrow. I just had a modest OC on the 780. I never go to crazy with that stuff because I don't really understand the ramifications of doing it.
I will try to recreate it when the flag drops on Friday. Sorry I couldn't be more help.
Thanks Orange! A little bit of recognition ain't so bad.

 

Incognito II
Call me crazy, setup a GTX 560 to run the long work units yesterday as a test. Currently at 91% complete with an elapsed time of 16hrs 54mins showing a remain time of 2hrs 35mins. SO it appears atm it will complete in time for the bonus points. So is it worth the effort to use this old hardware?

According to your numbers, you're good to go, but remember to switch to short tasks the last 24hrs, just to be safe.
However.... GPUGRID WU's comes in different flavor's. Even if the one you tested was ok, the next wu might be different.
I'd say you could start of with the long ones, but keep an eye on it.
One crazy thing though. The screenie above..... when yodap did a great run with his 780, take a look at the guy above......

Do we know if these teams are bunkering or not?  Does the bonus system discourage it? Is this our first time in the monthly contests?

yodap
Do we know if these teams are bunkering or not?  Does the bonus system discourage it? Is this our first time in the monthly contests?

Some crunchers bunker, and they are easy to spot.
I beleive bonus system disencourage it because to receive bonus, you must Return the Wu within a certain time.
I SAY NO BUNKER. We don't have to. We are strong enough as it is.
I also think this is Our first time. I checked the Challenge history, and couldn't find us.
Actually a bit strange, since we have some GPU firepower to show.
I might be wrong, but if all of us hard as we can here, we might have a fun fight, possibly about Third Place.
There are 17 teams now, and one of the Challenges before had 170 participants, so don't expect to be lonely
 
If I was on another team, and saw us, all signed up, I'd think... OH NO... Not THEM.
 
Get ready to rumble boyz

Thanks Mr Orange!
I'm against bunkering unless encouraged by the contest creator. 
I'll be running long runs for as long as possible on 970, 780, 660ti(x3) and a 650ti boost. I know our small but growing team will shine because we care!
Jumping in after 4:00 PM EST!

I'll be up and running by this evening.  One 780 Classy (long runs) and a 560ti (short runs).

If all goes well, I will be up and running tonight 2 x 780 SCs.
rs

So your going to be recommending against cpu tasks? Or do you just plan to put the cpu to better use elsewhere?
 
EDIT: Also, just clocked a long task at 5:40 with the cpu running a task on 12 cores. If anyones not running a task on the cpu with a gtx 980 @ about 1450mhz core can you see if your about the same times?

Orange_1050
 
In fact, I don't recommend anything regarding what wu's to run. That will be something for all to decide for themselves.
But actually, I was thinking of running SIMAP along With this GPUGRID Challenge since SIMAP is Our second POTM this month.
BUT... if we are struggeling hard and very even against another team, I'll do whatever in my Powers to push all the point I can, and that's when those CPU WU's might be called in

Looks like I have some testing to do then. I wouldn't mind running these cpu units if they would just not always demand all of the cpu. 

My GTX 680 and GTX 780 Ti have been crunching GPU Grid all week. I don't think I'm going to push the overclocking on either one since I am pushing into the area where heat problems start to happen. So I'm about as wound up and ready to rumble as I'm going to get.

God start! Khali Geos52 Devlin and Mflanaga posting results. Khali is in 5th.
Eagle07 and brilong are leading [H] to first. Maybe brilong is crunching on new card
My first long runs will drop in about 15 min.