1) Message boards : Number crunching : Odd credit behaviour (Message 39925)
Posted 9 hours ago by Profile Retvari Zoltan*
http://technicaldifference.net/mikeiv/GPUGridStats.png

This could happen if you merge your retired hosts to your new hosts, and the new ones receive the old hosts' credits.
However they should have the same CPU at least, and there are some other prerequisites for successful merging.
2) Message boards : Graphics cards (GPUs) : Upgrade to a Titan Z? (Message 39830)
Posted 1 day ago by Profile Retvari Zoltan*
Dual GPU-ed graphic cards like the GTX 295, GTX 590, GTX 690, GTX Titan Z have the cooler fan between the two GPU chips, in the middle of the card. Therefore the card emits its heat evenly in its both ends, so the half of that heat remains inside the computer's case, which could result in overheating, if the case fans aren't arranged accordingly (there should be fans in the front of the case which blows the hot air out, and fans on the side panel blowing cool air in). So the adequate case fan arrangement is different for dual GPU-ed graphic cards and for single GPU-ed graphic cards. Mixing them in a single host needs a little experimentation to find the right airflow direction for the best cooling.
I'm sure that at some point there will be a dual GPU-ed version of the GTX980, and that will much better bang for the buck than a Titan Z. If there will be a GTX version of Big Maxwell GPU (like the GTX 780 Ti of the Big Kepler GPU), it will be the best buy for GPUGrid.
Dual GPU-ed cards process two workunits at the same time, so a single but faster GPU-ed card is always the better choice.
3) Message boards : News : WARNING/CHALLENGE: VERY LONG WU (VERYLONG_CXCL12_confAna) (Message 39829)
Posted 1 day ago by Profile Retvari Zoltan*
There's another longer than usual batch called GERARD_CXCL12_confAnaFX_ada2, it took 9 hours for my GTX780Ti to complete.
It's granted 251,400 credits (including the 50% bonus).
4) Message boards : News : WARNING/CHALLENGE: VERY LONG WU (VERYLONG_CXCL12_confAna) (Message 39802)
Posted 2 days ago by Profile Retvari Zoltan*
ERRRRRRR. 217,000 secs of compute for nothing. It says error while computing. That was a waste. Yes, I am frustrated. I know I shouldn't be but 60 hours of compute down the drain. Don't know why. Any ideas?
http://www.gpugrid.net/result.php?resultid=13744228

This task's log contains a lot of error messages:

# The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 36378000) ... SWAN : FATAL : Cuda driver error 715 in file 'swanlibnv2.cpp' in line 1965. # SWAN swan_assert 0 ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 37079000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 37133000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 37963000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 41979000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 42124000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 42709000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 42718000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 43071000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 44636000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 44846000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 45046000) ... # The simulation has become unstable. Terminating to avoid lock-up (1) # Attempting restart (step 48215000) ... # The simulation has become unstable. Terminating to avoid lock-up (1)

Probably it's overclocked, and it can't take that much of that.
Its temperature doesn't go over 73°C, so it's not overheating.
You should either decrease the GPU clock by 10-20-50MHz, or increase the GPU voltage by 25mV (increase its power target).
But I don't recommend the latter.
5) Message boards : Wish list : Screensaver (Message 39795)
Posted 2 days ago by Profile Retvari Zoltan*
Lacks the graphics! need a screensaver, it would be interesting to see something to work on computers. Will realized graphics in your project?

There's no screensaver, but you can visualize the molecules your computer is working on.
See this post.
6) Message boards : Wish list : Screensaver (Message 39794)
Posted 2 days ago by Profile Retvari Zoltan*
It has been discussed before. Unfortunately no, because as far as I remember it would require us to ditch BOINC and write our own software and we don't have the manpower to develop something like this which would be of no scientific merit.

I agree that there are more important things in this research you (we) do than making a screensaver, when the project's staff lacks manpower.
You don't have to ditch BOINC, as many other projects have their own screensaver. A little modification of your current application is surely necessary though.
Have you ever thought of that if there would be an eye catching screensaver it could lure more crunchers (and perhaps donors) to this project? It's clear that a screensaver directly has no scientific merit, but indirectly (by it's effect on the public) it does.
So if you would sacrifice some of your manpower for a short period (by delaying a publication) and make an eye catching screensaver, it could be beneficial in the long term. I'm not asking you to do it, because it's not the only way. You could ask on the forum for someone who would do a screensaver for you. Maybe there's now no one who can do it, but in the long term it could change. But first, your point of view on a screensaver should change.
7) Message boards : GPUGRID CAFE : The milestone thread (Message 39784)
Posted 3 days ago by Profile Retvari Zoltan*
Trotador, Dave, skgiven, dskagcommunity, Finrond, SEARCHER, John C MacAlister

Congratulations!
8) Message boards : Number crunching : Tip: How to obtain zero GPU buffer size and normal CPU buffer size (Message 39733)
Posted 4 days ago by Profile Retvari Zoltan*
So it is not quite what I am after in trying to maintain a low GPU cache but a normal CPU cache.

You can install two separate instance of BOINC manager on a single host, one for low cache setting and the other for normal cache setting, and then add the projects accordingly.
All you have to do is set the advanced options when installing the second copy of BOINC manager, and designate a different program and data folders (for example: add a "2" to the end of their names).
9) Message boards : News : WARNING/CHALLENGE: VERY LONG WU (VERYLONG_CXCL12_confAna) (Message 39695)
Posted 6 days ago by Profile Retvari Zoltan*
I've received another one of these, and I confirm that the upload size has been doubled, so this problem is fixed by now.
<file_info> <name>8x3-GERARD_VERYLONG_CXCL12_confAna-0-1-RND4073_2_8</name> <nbytes>0.000000</nbytes> <max_nbytes>256000000.000000</max_nbytes> <generated_locally/> <status>0</status> <upload_when_present/> <url>http://www.gpugrid.org/PS3GRID_cgi/file_upload_handler</url> </file_info> <file_info> <name>8x3-GERARD_VERYLONG_CXCL12_confAna-0-1-RND4073_2_9</name> <nbytes>0.000000</nbytes> <max_nbytes>512000000.000000</max_nbytes> <generated_locally/> <status>0</status> <upload_when_present/> <url>http://www.gpugrid.org/PS3GRID_cgi/file_upload_handler</url> </file_info>

However it is not clear for me, why is the size of the _9 file bigger than the size of the _8 file, while the in the reality it's the opposite.
10) Message boards : News : WARNING/CHALLENGE: VERY LONG WU (VERYLONG_CXCL12_confAna) (Message 39661)
Posted 7 days ago by Profile Retvari Zoltan*
My second very long workunit is uploading.
7x6-GERARD_VERYLONG_CXCL12_confAna-0-1-RND0829_0


Next 10