Posts by petri33 - 3(2MB+1AP) tweaking


log in
1) Message boards : Number crunching : The Highest Ranked SETI AMD Host is a MAC: Time for a STOCK MAC APP? (Message 1646306)
Posted 1 day ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
It works!

...



Nice work!

If You have problems with high cpu usage try googling libsleep or take a look at https://bitcointalk.org/index.php?topic=180658.0

Google "nvidia sched yield" too and find the nvidia initialization code from your seti app source. Try with busy waiting/sleep/yield or whatever options Mac has.
2) Message boards : Number crunching : The Highest Ranked SETI AMD Host is a MAC: Time for a STOCK MAC APP? (Message 1645795)
Posted 3 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
Please try : rename it so that it has numbers.
3) Message boards : Number crunching : The Highest Ranked SETI AMD Host is a MAC: Time for a STOCK MAC APP? (Message 1645762)
Posted 3 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
You need to have a file MultiBeam_Kernels_r2120.cl in your seti directory.

Just copy the MultiBeam_Kernels.cl file to that file.
4) Message boards : Number crunching : GPU performance of GTX580 and GTX780Ti (Message 1644728)
Posted 6 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
Thank you for your advices.
I quickly appended "app_config.xml" to run 2 tasks per GPU.

But why 2 or 3 tasks are recommended for the performance?
BOINC's default setting is 1 task per GPU.



One at at time works for all (If any).

Sending work, receiving answers, seding new .. They all have a wait time (latency). Both the CPU and GPU can do something valuable when waiting for the other.
5) Message boards : Number crunching : Error While Computing.. (Message 1644660)
Posted 6 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
... and the stdrerr is:
[]
<core_client_version>7.2.42</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code -4 (0xfffffffc)
</message>
<stderr_txt>
setiathome_v7 7.00 DevC++/MinGW/g++ 4.5.2
libboinc: 7.1.0

Work Unit Info:
...............
WU true angle range is : 2.714941
Optimal function choices:
--------------------------------------------------------
name timing error
--------------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrumUnrolled 0.001067 0.00000
Memory allocation failed in ChooseChirp
BH SSE folding 0.001108 0.00000
SETI@home error -4 Can't allocate memory
!binsAboveThreshold
File: ../pulsefind.cpp
Line: 89


</stderr_txt>
]]>
[]
6) Message boards : Number crunching : GPU performance of GTX580 and GTX780Ti (Message 1644657)
Posted 6 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
Do you use app_config.xml to run 2 tasks per GPU?
... ...


I'd too advise of doing 2 (maybe 3) at a time.
7) Message boards : Number crunching : GPU performance of GTX580 and GTX780Ti (Message 1644656)
Posted 6 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
I can't comment on the gtx580, but my pair of GTX780 are doing about 45000....

My three GTX780 are doing ~ 70000+ (on MB only for now).
8) Message boards : Number crunching : near vlar? (AR 0.01) super slow on NVIDIA gtx 780 (Message 1642245)
Posted 12 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
The vlars can be sent to my nvidia any time from now on. I made a bug to the code that checks the AR and assigns a char zero to null whenever ar < z.

setiathome enhanced x41zc, Cuda 6.50 special Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 128k elements. Work Unit Info: ............... WU true angle range is : 0.008647 SIGSEGV: segmentation violation Stack trace (8 frames): ../../projects/setiathome.berkeley.edu/setiathome_x41zc_x86_64-pc-linux-gnu_cuda65[0x4ae0a0] /lib/x86_64-linux-gnu/libpthread.so.0(+0xfc90)[0x7f2cc7e6bc90] ../../projects/setiathome.berkeley.edu/setiathome_x41zc_x86_64-pc-linux-gnu_cuda65[0x41654e] ../../projects/setiathome.berkeley.edu/setiathome_x41zc_x86_64-pc-linux-gnu_cuda65[0x423275] ../../projects/setiathome.berkeley.edu/setiathome_x41zc_x86_64-pc-linux-gnu_cuda65[0x42def4] ../../projects/setiathome.berkeley.edu/setiathome_x41zc_x86_64-pc-linux-gnu_cuda65[0x406a26] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7f2cc4618ec5] ../../projects/setiathome.berkeley.edu/setiathome_x41zc_x86_64-pc-linux-gnu_cuda65[0x40a027] Exiting... </stderr_txt> ]]>
9) Message boards : Number crunching : near vlar? (AR 0.01) super slow on NVIDIA gtx 780 (Message 1641731)
Posted 12 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
Thamk You Jason,

I'll check out how the Lunatics shortened wu's are done. That is just what I need. To test out with maximum size wu that nvprof can use.
10) Message boards : Number crunching : near vlar? (AR 0.01) super slow on NVIDIA gtx 780 (Message 1641324)
Posted 13 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
Thank You Zalster.

I knew that, but had had no time to check that the were actual vlars. And a real vlar (when processed) slows down the other tasks running on the same GPU at the same time.

Now I'd like to know how to make a shortened version of a vlar that I captured so that I can profile it under nvprof. If I run it as it is it generates too much profiling data.

Joe? Raistmer? Any?

How to make a shortened wu?

And then there remains a question why did I get some vlars for my NV..
11) Message boards : Number crunching : near vlar? (AR 0.01) super slow on NVIDIA gtx 780 (Message 1641224)
Posted 13 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
I've had a few MB tasks that run about 5000 seconds when running 3 at a time. The normal run time for a multibeam task is 450-1100 seconds when running 3 at a time.

The slow ones are 'near' vlars I guess?
12) Message boards : Number crunching : The Highest Ranked SETI AMD Host is a MAC: Time for a STOCK MAC APP? (Message 1640983)
Posted 14 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
Just a peanut gallery observation while taking a break:

any idea why your environment changed (first error example from xml_util.h)
c[1] = ( (bin[offset]>>6) | ((bin[offset+1]<<2)&0x3f) ); // 2+4
, which is perfectly fine C-Code, into
c[1] = (bin[offset]>>6) | ((((bin[offset+1]<<2)&0x)3)f); // 2+4

which is not ?

If the file on disk has that second code, then something broke the sources. or do the current repository sources have that damage ?

Good question. I can only guess. Using Xcode sometimes it will offer to 'fix' errors for you. All you do is double click on the offer and it will 'fix' it for you. That could have been what happened. I just downloaded the file from here;
https://setisvn.ssl.berkeley.edu/trac/browser/branches/sah_v7_opt/AKv8/db/xml_util.h, shut down Xcode and swapped the files and now that error is gone. I tried that with db/db_table.h but it didn't work. Right now the only Error left is;
/Users/TBar/mac_seti/AKv8/db/db_table.h
/Users/TBar/mac_seti/AKv8/db/db_table.h:331:20: Expression is not assignable
/Users/TBar/mac_seti/AKv8/client/chirpfft.cpp:48:10: In file included from /Users/TBar/mac_seti/AKv8/mac_build/../client/chirpfft.cpp:48:
/Users/TBar/mac_seti/AKv8/client/seti_header.h:43:10: In file included from /Users/TBar/mac_seti/AKv8/mac_build/../client/seti_header.h:43:

I did "double click" a few times with the old Xcode, apparently those are not showing as a problem. As soon as I can figure out how to save all my settings, I'm going to replace all the files with my zipped copy and work from there. Being only 1 Error away from a build, I will definitely build a copy first. Now how to 'Fix' this last problem?
if (where.size() = 0) return false;

if (where.size() == 0) return false;
13) Message boards : Number crunching : Lunatics Experience (Message 1640850)
Posted 14 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
The Symantec antivirus may be scanning your boinc and seti directories and the executable in memory and its file writes/reads etc.
14) Message boards : Number crunching : >1 Task per CPU? (Ans: Probably Not) (Message 1639519)
Posted 17 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
I think there is a ncpus or something in cc_config for testing purposes ..
15) Message boards : Number crunching : Show and tell your machine. Here's mine. (Message 1635858)
Posted 26 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
My GTX 780's have blown their fans out months ago. They're "passive" cooled now.

See pictures below:

My Rig
(currently running on spare part mobo P9X79-WS, while P9X79WS-E is being in service)


These are not currently used: (Side panel fans are not powerful enough, One Gtx780 GPU does not fit on mobo)




This is The Fan.




… blowing less than an inch away from the 3 GPUs (Their heatpipes recognizable)



What is not visible? -- The Fan is taped up almost completely from the GPU side. It has an opening for the air flow just where the GPUs are. From the fan input (air in) it is taped at the outer rim for to keep the fast moving air inside the fan: The only escape for the air is just where it is supposed to be -- towards the GPU section.
16) Message boards : Number crunching : Main cruncher dead (Message 1635291)
Posted 27 days ago by Profile petri33 - 3(2MB+1AP) tweakingProject donor
I made a mistake with the motherboard, should've gone with a gaming one instead of overclocking. It only supports 2xSLI, and those 2 slots are right next to each others. The 3rd slot can't be used as it is x4 and SLI requires a minimum of x8.

I already have 4 fans and a radiator on top in push/pull configuration, 200mm front intake, 200mm side intake, 140mm rear exhaust. There's no way to keep the upper GPU fans from taking the hot air from the lower GPU. I can try and use the second GPU in the 3rd slot and forget about SLI when I am not gaming, and see if the temps are manageable. We'll see, it depends on the RMA. If I don't get one, I am back to a single gpu setup anyway.



Hi,

I live here up in the north (near the Arctic circle).

Heat is a problem here too (at least in my computer room...)

My GTX 780's have blown their fans out months ago. They're "passive" cooled now.

See pictures below:

My Rig
(currently running on spare part mobo P9X79-WS, while P9X79WS-E is being in service)


These are not currently used: (Side panel fans are not powerful enough, One Gtx780 GPU does not fit on mobo)




This is The Fan.




… blowing less than an inch away from the 3 GPUs (Their heatpipes recognizable)




... And yes. Check The 24 pin connector to the MoBo as some one stated right at the beginning of this thread. Look at the cable pins near the top of your motherboard - mine have melted (plastic burned away, copper 'tubes' visible). Stability restored by connecting a molex extra power cable to the MoBo.


p.s. I remember someone asking for pictures of my system -- Well, here they are. In a totally obscured and a wrong place, but anyway.
17) Message boards : Number crunching : HW donations (Message 1620046)
Posted 29 Dec 2014 by Profile petri33 - 3(2MB+1AP) tweakingProject donor
I'll give up some old hardware:

1) a 560 Ti and
2) a GTX 660.

Both of them have cooled off now for more than a year on the shelf. They might need some lubrication.

You may opt to have just one but I guess the shipping costs are about the same for both of them.


Hey petri33, I will take them both.

Look at Your PM.


WezH is now the lucky one. He'll be receiving a packet soon after New Year.
18) Message boards : Cafe SETI : Word Link Discussion Thread III (Message 1618596)
Posted 25 Dec 2014 by Profile petri33 - 3(2MB+1AP) tweakingProject donor
I wish me a crappy christmas.
19) Message boards : Cafe SETI : Word Link #76 (Message 1618588)
Posted 25 Dec 2014 by Profile petri33 - 3(2MB+1AP) tweakingProject donor
non-test
20) Message boards : Cafe SETI : Beet's Give Us a Caption #58 (Message 1618585)
Posted 25 Dec 2014 by Profile petri33 - 3(2MB+1AP) tweakingProject donor
No snow!


Next 20

Copyright © 2015 University of California