Posts by petri33


log in
1) Message boards : Number crunching : What's the best valued GPU? (Message 1761138)
Posted 11 days ago by Profile petri33
My GTX 980's report each eating 140-150W power @ 1310MHz and ram at 7200.
I use nvidia-smi to get the card status.

While being quite expensive they do give some decent RAC.

The three 970's that you can find on the top hosts page seem to be doing quite well also.
2) Message boards : Number crunching : Your kooky computers (Message 1760842)
Posted 12 days ago by Profile petri33
UNTIL YOU UNHIDE YOUR PC(S) NO ONE CAN OR WILL HELP YOU: "Computers hidden".
My suspicions are that you have installed one of the customized apps (aka "Lunatics Installer").

Sorry to disappoint you but I don't know what that is although I have seen it mentioned somewhere. And I also don't know why a customized installer would be bad. Or why it matters to anyone but me what computers I have. You can see what projects I have worked on. At least I think you can. So who cares what my computers are? And they say you should restrict whatever personal data you can online, so I chose to hide that for that exact reason, believe it or not. 8) I have also been volunteering computers for seti@home since... I'm not sure, but I got in trouble at work for it at the time, but it was attached to a work email that I dont have access to now. But I think you can see I have credits in the low thousands, not millions or kajillions like you guys, so I have no computer you'd even consider using as your cat's litterbox. 8)


Restricting your personal information is ok (since it is allowed to hide it). If a seti program crashes it can reveal your username by listing your user path in error report. I have seen that happen.
3) Message boards : Number crunching : I'm Trying to Build an OSX CUDA App... (Message 1756078)
Posted 29 days ago by Profile petri33
If you're referring to my files, I think I sent you two versions of each.

ref-...res - that is the actual result file that would be uploaded to the project servers: it is the authoritative, full-precision file. It's a simple text (xml) file, but it's written with *nix line-endings, which can make it hard to read on Windows. Use either WordPad or NotePad++.

ref-...summary.txt - that is a simpler summary file, designed to make it easier to compare a result visually when all you have to compare with is the std_err of an OpenCL application. I think I sent you the summariser tool as well: the ReadMe explains that the figures are truncated (not rounded) to six decimal places.

For predicting normal validation, the 6-figure summary is probably enough, but for application checking, you should also refer to the Q-value reported by the rescmpv5 tool, which works on the full-precision files.

I don't think the OpenCL apps list the triplet periods in std_err, so I didn't bother summarising it. For that, you'll need to look in the full result file.


Hi Richard,

I'm running under Linux. I use your (thank you) full precision ref-res files that I copied over my ref-result.app-name.PGxxxx.wu xml-style files in testData and renamed them so that my old rescmp5_l can do the comparison and Q-value reporting every time I test a new app. Current Q is 99.17% - 99.97% depending on wu.

I can read the result files with emacs in extremely readable pretty printed screenlayout and do a diff 2 buffers and browse for next difference hitting 'n'.
Precision seems pretty good with current version.

Petri
4) Message boards : Number crunching : I'm Trying to Build an OSX CUDA App... (Message 1756057)
Posted 29 days ago by Profile petri33
FYI,

I'm testing now against the reference-results I got from Richard. Looking better every day. My machine makes almost no errors (one bad for every 400 000 points), The number of invalids is dropping.

Jason and TBar may want to check their mailbox for latest source.

One question: Where is the period for triplts and pulses calculated and possibly truncated/rounded for output?

REF: <period>0.9961472</period>
MY : <period>0.99614721536636</period>
5) Message boards : Number crunching : How much does SETI@home affect your computer performance long-term? (Message 1755033)
Posted 9 Jan 2016 by Profile petri33
PSU's die. That's a fact. 7 years warranty helps. Keep loyal to your supplier.
www.jimms.fi is mine.
6) Message boards : Number crunching : New app? (Message 1754919)
Posted 8 Jan 2016 by Profile petri33
So far I can report 63 consecutive valid tasks on my GTX 750 using the r3314 app:
http://setiathome.berkeley.edu/host_app_versions.php?hostid=7563243

It runs well but it seems to put a lot of stress on the hardware, even with use_sleep. It just took me about 20 to 30 seconds to switch from the Boinc manager to the desktop. I thought it would crash every second but it's still running. So, with the CPU at 100% the system seems to be on the edge.

Can't say much about performance yet. The GFLOP rating is lower then for the V7 app, but the doesn't have to mean anything.

So, is there already a newer version then the one I use? Would love to test further and see if the (un)responsiveness improves. Or maybe somebody has suggestions for a commandline?


I have no time to see if this has been answered, but "use_sleep" might help.
I'm going to put my 6 & 8 yr to bed and read a good night story --- "use sleep".
7) Message boards : Number crunching : Dumb Question(s) About Seti v8 (Message 1754212)
Posted 5 Jan 2016 by Profile petri33
So, as I understand what you said, before app_config.xml, there was NO way to run > 1 WU per GPU on stock?

I did not know that.

Not at SETI. Einstein@Home has a bespoke solution, but for their own project only.


bespoke == tailored to suit (ones needs)
8) Message boards : Number crunching : Seti@home v8 (Message 1753843)
Posted 4 Jan 2016 by Profile petri33
Except there ISN'T any Version 8 CUDA Source code out yet, making it Impossible to have a Version 8 CUDA App.
https://setisvn.ssl.berkeley.edu/trac/browser/branches/sah_v7_opt
Since I'm waiting to build My own Version 8 CUDA App, I know of such details.


Well there is a version in your mailbox.
I'm waiting for the first result to either verify, fail or something.

http://setiathome.berkeley.edu/workunit.php?wuid=2024980992

I got the email, but the attachment is missing?
I have only the text, no files....
:-(


Now you have it. My bad. Sorry.

Tried compiling in Yosemite but I'm getting the same Error in a number of sections;
/usr/include/dispatch/object.h(143): error: expected an identifier /usr/include/dispatch/object.h(362): error: identifier "dispatch_block_t" is undefined

Because of those errors it ends with;
clang: error: no such file or directory: 'cudaAcceleration.o'
clang: error: no such file or directory: 'cudaAcc_CalcChirpData.o'
clang: error: no such file or directory: 'cudaAcc_fft.o'
clang: error: no such file or directory: 'cudaAcc_gaussfit.o'
clang: error: no such file or directory: 'cudaAcc_PowerSpectrum.o'
clang: error: no such file or directory: 'cudaAcc_pulsefind.o'
clang: error: no such file or directory: 'cudaAcc_summax.o'
clang: error: no such file or directory: 'cudaAcc_transpose.o'
clang: error: no such file or directory: 'cudaAcc_utilities.o'
clang: error: no such file or directory: 'cudaAcc_autocorr.o'
make[2]: [seti_cuda] Error 1 (ignored)

Later today I'll try it in Mountain Lion.


See if this helps ...
http://stackoverflow.com/questions/27976312/how-to-cope-with-non-gcc-compatible-code-in-os-x-yosemite-core-headers/28014302#28014302
9) Message boards : Number crunching : Show and tell your machine. Here's mine. (Message 1753731)
Posted 3 Jan 2016 by Profile petri33
Try this calculator to find out how much PSU you need
http://outervision.com/power-supply-calculator

There's an EVGA one that I like also but their website is down until next week.


Yep, I know that calculator, and I did use it before installing third Ti. And it always calculated under 500W PSU.



My 980's are using 147W max according to nvidia-smi. The 780's do not show their wattage.
10) Message boards : Number crunching : Seti@home v8 (Message 1753730)
Posted 3 Jan 2016 by Profile petri33
My GPU's did not find 30 autocorrelations with these:

http://setiathome.berkeley.edu/workunit.php?wuid=2025281218
http://setiathome.berkeley.edu/workunit.php?wuid=2025281338
http://setiathome.berkeley.edu/workunit.php?wuid=2025281344

Are there?
11) Message boards : Number crunching : Seti@home v8 (Message 1753654)
Posted 3 Jan 2016 by Profile petri33
Except there ISN'T any Version 8 CUDA Source code out yet, making it Impossible to have a Version 8 CUDA App.
https://setisvn.ssl.berkeley.edu/trac/browser/branches/sah_v7_opt
Since I'm waiting to build My own Version 8 CUDA App, I know of such details.


Well there is a version in your mailbox.
I'm waiting for the first result to either verify, fail or something.

http://setiathome.berkeley.edu/workunit.php?wuid=2024980992

I got the email, but the attachment is missing?
I have only the text, no files....
:-(


Now you have it. My bad. Sorry.
12) Message boards : Number crunching : Seti@home v8 (Message 1753645)
Posted 3 Jan 2016 by Profile petri33
< 80 secs to crunch... is that right?


Yes. Shorties are 66-70 seconds. Jus as they were with v7.
13) Message boards : Number crunching : Seti@home v8 (Message 1753632)
Posted 3 Jan 2016 by Profile petri33
That's a special case :)


Very special 75 minutes old.

http://setiathome.berkeley.edu/workunit.php?wuid=2024687465
14) Message boards : Number crunching : Seti@home v8 (Message 1753618)
Posted 3 Jan 2016 by Profile petri33
Except there ISN'T any Version 8 CUDA Source code out yet, making it Impossible to have a Version 8 CUDA App.
https://setisvn.ssl.berkeley.edu/trac/browser/branches/sah_v7_opt
Since I'm waiting to build My own Version 8 CUDA App, I know of such details.


Well there is a version in your mailbox.
I'm waiting for the first result to either verify, fail or something.

http://setiathome.berkeley.edu/workunit.php?wuid=2024980992
15) Message boards : Number crunching : Panic Mode On (101) Server Problems? (Message 1753429)
Posted 3 Jan 2016 by Profile petri33
Thank You!



A square root. Right to left.
16) Message boards : Number crunching : Seti@home v8 (Message 1753394)
Posted 2 Jan 2016 by Profile petri33
As far as Eric and the rest of the project are concerned, there's no such beast as "GPU work". There's just work, and they couldn't care less what tool is used to crunch it - you can use an abacus if you like, provided it produces accurate results.

The major purpose of the new v8 application was to gain the ability to process data recorded at the Green Bank Telescope, and other telescopes in general. There isn't much of that yet, so it makes sense to start slowly - testing that SETI's specialist recorder, data transfer, splitters, volunteer processing, validator, assimilator, and science database can all handle the new work.

All those components have to be in place before the entire pathway can be tested under stress. And, I'm sorry to say, volunteers' GPUs seem to be the greatest source of random errors in the system. It makes sense to start with CPU applications only: their relative simplicity and generally low error rate will make any errors anywhere else in the chain stand out more clearly above the noise.


Plus all the work a lot of people have done behind the scenes. Things do not just appear. The world does not have an intelligent design behind it. But Seti@home has.
17) Message boards : Number crunching : New app? (Message 1753381)
Posted 2 Jan 2016 by Profile petri33
@ Fawkesguy,

Look at the line immediately following

<app_version>

In every case, it's an <app_name> - EXCEPT the one you added in red. It'll be

<app_name>setiathome_v8</app_name>

of course.

@ TBar - he quoted your post, which is where the error came from. Please proof-read, or better yet test, before posting.



Richard, thank you for catching that. :-)


I've now got the v8 app and app_info set up on all of my machines, but it seems that only my Windows machine is grabbing v8 WU's. Any idea what might be causing that?

EDIT: Found a typo in my app_info. All good now. :-)

Yes those typos can elude many. I wouldn't thank Richard too much, all he did was look at the changes I had made at Beta 30 minutes earlier, http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2266&postid=55569#55569, and then decide it would be a good time for some ragging. Strange he didn't see the problem until I had posted it.

After a couple days I now have Version 8 GPU Apps for ALL my machines. It was kinda hectic, but now it's over. Time for a little R&R.


Congrats! Running, awesome!

R & R : as of Research and Rebuild?

7:30 min for a 2.70 ar with v8 (36.84 credits) and
2:51 min for a 2.59 ar with v7 (40.74 credits).

I know. You have an advanced version of v7.

Just wait.
18) Message boards : Number crunching : Panic Mode On (101) Server Problems? (Message 1753333)
Posted 2 Jan 2016 by Profile petri33
I'm on CPU now, both v7 and v8. Mixed use is a good thing despite of any warnings.
No GPU tasks for v7. The cache is empty.

I'll go to sleep. Dream of jason_gee's CUDA update and wake up to a busy tomorrow poring a v8 CUDA app for linux.

That's my dream.

Petri <3 <3
19) Message boards : Number crunching : New app? (Message 1753318)
Posted 2 Jan 2016 by Profile petri33
A quick look at the -9's tells me that different codepaths were chosen and still the result validated. Both determinded the same result. My quess is that it is a valid nonvalid.
20) Message boards : Number crunching : New app? (Message 1753316)
Posted 2 Jan 2016 by Profile petri33
Thank you for everyone involved in developing the new v8 app.

I have run some WU's and all have validated.
I run on linux 64-bit with i7 3930K. The app I'm using is 32 bit.

Which version would be best suited for my GPU?

See run times here: http://setiathome.berkeley.edu/results.php?hostid=7475713&offset=0&show_names=0&state=4&appid=29

Petri

Petri, I question some of your results as you are validating a -9 overflow against a -9 overflow. (of course that may be correct)


I have not had time to see the results with my brains, just eyes. Will look tomorrow. I guess the -9's are a thing of the app being new and finding untraced bugs. The buildmasters will sort them out eventually.

And I had a crash at my mahhine about an hour ago. There was a message saying an internal error happened. May be related to that.

Thanks,
Petri


Next 20

Copyright © 2016 University of California