Posts by Vipin Palazhi

21) Message boards : Number crunching : Panic Mode On (81) Server Problems? (Message 1335430)
Posted 107 days ago by

None of my machines are able to report back to the servers, all requests are timing out. Is there any news from the seti staff regarding this prolonged issue?

I will try abusing the retry button, something which I have never done before.

22) Message boards : Number crunching : Panic Mode On (81) Server Problems? (Message 1335154)
Posted 108 days ago by

Vipin Palazhi

I cant use any proxies as they are all blocked by the ISP here. So I just need more patience till the issue gets resolved.

23) Message boards : Number crunching : Panic Mode On (81) Server Problems? (Message 1335149)
Posted 108 days ago by

Vipin Palazhi

Super crunchers like Juan's will go through their cache of work quickly compared to medium machines. Even the modest rigs that I run are all dry. Btw, does anyone have an idea what is going on in seti labs and when we can expect the work flow to resume? No update to be seen anywhere regarding the situation .

24) Message boards : Number crunching : Added GTX670 Not getting expected performance Boost (Message 1335118)
Posted 108 days ago by

Vipin Palazhi

The current issues with the server and unavailability of enough WUs to fill the cache will certainly impair your system from delivering its maximum potential. You should see better RAC once the WUs start flowing freely. I am using 6.10.60 as well and it shouldnt affect performance as long as you have the right optimized apps. You may want to upgrade your Lunatics MB file to x41zc and set up your GPU to crunch astropulse. The following links should help.

x41zc
Astropulse
You can download the latest files from Arkayn's website

I will let the more experience crunchers explain the gflops of the GPU.

25) Message boards : Number crunching : A Modest Proposal for Easy FREE Bandwidth Relief (Message 1334754)
Posted 109 days ago by

Vipin Palazhi

Thanks for that link bill. However, that means the servers are still located within the berkeley network, clogged by all the politics going on. I do remember reading someone offering to host some of the servers for free. Now, I am no expert in this area, just thinking out loud.

Apologies if the above point has already been discussed.

26) Message boards : Number crunching : A Modest Proposal for Easy FREE Bandwidth Relief (Message 1334744)
Posted 109 days ago by

Vipin Palazhi

Has anyone considered the options of splitting and distributing the work units from a location other than berkeley? It will then redirect a good amount of 'outbound' data away from these overloaded berkeley bandwidth. The completed WUs will only be reported back to berkeley.

27) Message boards : Number crunching : Panic Mode On (81) Server Problems? (Message 1334590)
Posted 110 days ago by

Vipin Palazhi

I have been crunching for a few years now without ever actually bothering about how long it takes for the discs to be recorded and how/when it gets transported from Arecibo to Berkeley.

28) Message boards : Number crunching : OpenCL apps are available for download on Lunatics (Message 1333584)
Posted 112 days ago by

Vipin Palazhi

We didn't see such big differencies in offline tests. Most probably you changed app config also. There is almost no changes in AP from r1363

All I did was replace the exe file and edit the file name in the app_info. Is there any reason why the GPU isn't getting fully loaded?

29) Message boards : Number crunching : OpenCL apps are available for download on Lunatics (Message 1333537)
Posted 113 days ago by

Vipin Palazhi

I have updated my rig with r1761. It is now running two units right now on a GTX480 and my observations are:
1. The GPU load is all over the place ranging from 15 - 95%
2. The crunching time seems to have doubled from an average of 65 minutes for r1363

Both have 1 CPU core allocated to it.

Edit: The completion time for the above WU was 2 hours 17 minutes.

30) Message boards : Number crunching : astropulse GPU (Message 1333158)
Posted 114 days ago by

Vipin Palazhi

Not really, My experience of running AP on my 9800GTX+ is that it's just not worth it,

Claggy

Thanks for the info Claggy.

31) Message boards : Number crunching : astropulse GPU (Message 1332847)
Posted 114 days ago by

Vipin Palazhi

Is it worthwhile running AP on NV 200 series GPU? Or is it only good for the fermi cards?

32) Message boards : Number crunching : Panic Mode On (80) Server Problems? (Message 1331490)
Posted 119 days ago by

Vipin Palazhi

Cannot report the completed tasks from this part of the world since the past couple of hours. All requests are timing out on all my rigs.

33) Message boards : Number crunching : Missing coprocessor (Message 1330696)
Posted 121 days ago by

Vipin Palazhi

I have had similar issues when I tried upgrading the boinc version on my rig from 6.10.60 to 7.0.28. I have been running nVidia driver 306.81 for a while now. Plus the issue is certainly not due to the processor as I have the newer AMD Phenom II X6 1055T. I have now reverted to 6.10.60 and all is well.

34) Message boards : Number crunching : astropulse GPU (Message 1330686)
Posted 121 days ago by

Vipin Palazhi

you can skip AP6_win_x86_SSE2_OpenCL_NV.pdb if it not will be listed in your app_info

Thanks Raistmer. I have now added all the files to the working directory and amended the app_info file. Started boinc and its crunching normally, however, no AP action from the GPU. Could you please clarify if the AP work units for nVidia GPU are different from the CPU ones or will the same units be taken up by the GPU? Also, does the boinc version have anything to do with this (I am running 6.10.60)

The <platform>windows_intelx86</platform> is also bothering me as I use AMD rigs.

35) Message boards : Number crunching : astropulse GPU (Message 1329158)
Posted 125 days ago by

Vipin Palazhi

"app_info section for NV OpenCL AP" posted by Claggy:
http://setiathome.berkeley.edu/forum_thread.php?id=70264&postid=1328639#1328639

! to retain formatting (indents) - do not copy from the page but click [Quote] under the post and copy from there

So all I need to do is append the given app_info section to the one I already have, create the ap_cmdline.txt and put the four files (AP6_win_x86_SSE2_OpenCL_NV_r1363.exe, AP6_win_x86_SSE2_OpenCL_NV.pdb, AstroPulse_Kernels_r1363.cl and libfftw3f-3.dll) in setiathome.berkeley.edu folder and restart boinc?

Am I correct in assuming that OpenCL files required are already present along with the latest nVidia drivers?

Will the information in <platform>windows_intelx86</platform> need any change for AMD systems, and must <plan_class>cuda_fermi</plan_class> be changed depending on the type of nVidia cards?

36) Message boards : Number crunching : Faster GPUs with Nvidia 310.70? (Message 1328603)
Posted 127 days ago by

Vipin Palazhi

1) Use the one that fits your processor, OS, and GPU. e.g. 64-bit, Win7, 200-series.
2) If I check for, e.g. 64-bit Win7, 2000 series on the Nvidia site it only offers one driver. Again, choose the one that suits your OS (32- or 64-bit).
Never install two drivers -- I doubt if you can anyway, the latest will always supplant the earlier.

@ Ivan
I am not sure why, but when I search for 64bit XP drivers for GTX 260, the only supported devices shown are GeForce GTX 275 and GeForce GTX 260. If I use GTX 285 in the search term, the devices shown are GeForce GTX 295, GeForce GTX 285, GeForce GTX 280, GeForce GTS 250, GeForce GTS 240, GeForce GT 230, GeForce GT 240, GeForce GT 220, GeForce G210, GeForce 210, GeForce 205. This is what lead me to the conclusion that, although the driver version is the same for both (310.90) there might be some differences in them.

I performed an automatic driver detection which also gave me two different download options. I do see that you are using the uk site, which is obviously giving a different result.

@ Claggy
I suppose cuda23 would be the best option for me at this moment. May I also ask what would be the best choice for my GTX480, as you mentioned that cuda5 app is the slowest?

37) Message boards : Number crunching : Faster GPUs with Nvidia 310.70? (Message 1328551)
Posted 127 days ago by

Vipin Palazhi

Now that the servers are all back online and all my rigs are getting work, I am trying to replace the x41g with x41zc. The rig has a GTX260 and a GTX285, and is running nVidia driver version 285.58.
1. Arkayn's site shows many versions of the cuda file and I am not sure which one to use?
2. The x41zc_Winx64_cuda50 requires driver version 301.48 and higher, which presents me with another issue. The nVidia site gives two different downloads for the latest version 310.90 for GTX 260 and 285 (the last unified driver was 301.42). Can these two versions be installed simultaneously or will it cause any clash?

38) Message boards : Number crunching : Panic Mode On (79) Server Problems? (Message 1311429)
Posted 170 days ago by

Vipin Palazhi

All my crunchers ran out of work during the outage and I switched them all off except one. If others too have done the same, then I guess the electricity consumption all over the world might have seen a dip during these days.

Good to see that things are up and running, however, all I am seeing is - Scheduler request failed: Failure when receiving data from the peer, and Scheduler request failed: Timeout was reached.

39) Message boards : Number crunching : -9 result overflow (Message 1311059)
Posted 177 days ago by

Vipin Palazhi

Well I tried restarting but no luck. One GPU (card 2) is going through the CUDA WUs in 15 seconds while the other (GPU 1) is crunching fine. I stuck this GPU in another rig and the display was all faded out. I am now guessing that the card itself is giving out. I have taken it 'off the production' line until I can figure this out over the weekend.

Eidt: The event log says 'no usable GPUs found'.

40) Message boards : Number crunching : -9 result overflow (Message 1311014)
Posted 178 days ago by

Vipin Palazhi

Double post, please ignore.

Previous 20 · Next 20