Now that v7 has rolled out..........

Message boards : Number crunching : Now that v7 has rolled out..........
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 11 · 12 · 13 · 14 · 15 · 16 · 17 · Next

AuthorMessage
Profile Vipin Palazhi
Avatar

Send message
Joined: 29 Feb 08
Posts: 286
Credit: 167,386,578
RAC: 0
India
Message 1374212 - Posted: 31 May 2013, 12:16:34 UTC
Last modified: 31 May 2013, 12:47:33 UTC

For some strange reason libfftw3f-3-3_upx.dll keeps getting deleted in all my rigs. Three of them are running AVG while my main rig has Bitdefender. I have put up an exclusion for this as well as the v7 main executable files, although neither gave any warnings. I thought the warnings were only for the exe file.

All the completed v7 WUs are now piling under validation pending.
______________

ID: 1374212 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34428
Credit: 79,922,639
RAC: 80
Germany
Message 1374213 - Posted: 31 May 2013, 12:21:47 UTC - in response to Message 1374156.  

Now that my GTX460 cruncher is on stock app for the first time, it downloaded its first Nvidea AP to crunch using opencl. However, GPU utilization is very low. For MB WUs, GPU utilization is typically 80-90% (for 1 v7 WU), but the AP was only showing ca 4% GPU utilization. I freed up one CPU core and this increased to 18% with occaisional bursts up to 100%. Freeing up a second CPU core made no difference. Is this normal?

The AP WU is currently 24% complete and has been running for 1.5 hours, so it is heading for about 6 hours to complete. Does this sound right for a GTX460 running at 857MHz?


Yes, you need to free a CPU core for astropulses.

You can speed things up.

Add the following params to your ap_cmdline.txt
-unroll 10 -ffa_block 6144 -ffa_block_fetch 1536.




With each crime and every kindness we birth our future.
ID: 1374213 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1374265 - Posted: 31 May 2013, 13:51:47 UTC

There doesn't seem to any stock linux app for nvidia, will there be an opti app on Monday?


Will Lunatics_x41g_linux64_cuda32 work with v7?
ID: 1374265 · Report as offensive
Profile Fred E.
Volunteer tester

Send message
Joined: 22 Jul 99
Posts: 768
Credit: 24,140,697
RAC: 0
United States
Message 1374280 - Posted: 31 May 2013, 14:09:37 UTC

Is there a clever way of freeing up one CPU core when the GPU is running an AP but when it is running MBs which do not need it? I do not want to loose 25% of my CPU RAC and cannot lways be present to baby sit the machine.

With your version of BOINC, you can do this with an app_config.xml file in your SETI data directory. Here's mine - set the gpu usage value to the values you need for your card. You don't need the max concurrent line - I use it to limit the number of AP's running at one time. Setting a value of one for the cpu usage will reserve one core but only while the AP task is running.


<app_config>
<app>
<name>astropulse_v6</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>.33</gpu_usage>
<cpu_usage>1.0</cpu_usage>
</gpu_versions>
</app>
<app>
<name>setiathome_v7</name>
<gpu_versions>
<gpu_usage>.33</gpu_usage>
<cpu_usage>.06</cpu_usage>
</gpu_versions>
</app>
</app_config>

Add the following params to your ap_cmdline.txt
-unroll 10 -ffa_block 6144 -ffa_block_fetch 1536.

Mike, will r1316 accept those command line switches? (question I asked in msattler's thread).

Another Fred
Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop.
ID: 1374280 · Report as offensive
Profile William
Volunteer tester
Avatar

Send message
Joined: 14 Feb 13
Posts: 2037
Credit: 17,689,662
RAC: 0
Message 1374335 - Posted: 31 May 2013, 15:13:47 UTC - in response to Message 1374265.  

There doesn't seem to any stock linux app for nvidia, will there be an opti app on Monday?


Will Lunatics_x41g_linux64_cuda32 work with v7?

Frankly we haven't the faintest idea.
It has AC code but it's not been alphaed on Linux.
update your app_info with the tags that have been published and give it a try?

Pay close attention to any problems.

x41g is going to have a higher rate of inconclusives, but it may just work.
A person who won't read has no advantage over one who can't read. (Mark Twain)
ID: 1374335 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51511
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1374346 - Posted: 31 May 2013, 15:24:18 UTC

Well, things seem to have settled a bit here.
My RAC drop is not so declined as before.
I have all the kibble I could wish for under the still enforced 100 task bit.

I suppose the drop in RAC is due to the fact that pendings are up a bit, and I am waiting for the new opti installer to do more than one task per GPU.

I have checked a couple of rigs, and it looks like I am at 90+ on all, so the few percent I am not using on a few rigs might account for something.
"Time is simply the mechanism that keeps everything from happening all at once."

ID: 1374346 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34428
Credit: 79,922,639
RAC: 80
Germany
Message 1374371 - Posted: 31 May 2013, 15:59:22 UTC - in response to Message 1374280.  
Last modified: 31 May 2013, 15:59:57 UTC

Is there a clever way of freeing up one CPU core when the GPU is running an AP but when it is running MBs which do not need it? I do not want to loose 25% of my CPU RAC and cannot lways be present to baby sit the machine.

With your version of BOINC, you can do this with an app_config.xml file in your SETI data directory. Here's mine - set the gpu usage value to the values you need for your card. You don't need the max concurrent line - I use it to limit the number of AP's running at one time. Setting a value of one for the cpu usage will reserve one core but only while the AP task is running.


<app_config>
<app>
<name>astropulse_v6</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>.33</gpu_usage>
<cpu_usage>1.0</cpu_usage>
</gpu_versions>
</app>
<app>
<name>setiathome_v7</name>
<gpu_versions>
<gpu_usage>.33</gpu_usage>
<cpu_usage>.06</cpu_usage>
</gpu_versions>
</app>
</app_config>

Add the following params to your ap_cmdline.txt
-unroll 10 -ffa_block 6144 -ffa_block_fetch 1536.

Mike, will r1316 accept those command line switches? (question I asked in msattler's thread).


With r_1316 you need to add/edit it in the cmdline in appinfo.xml.


With each crime and every kindness we birth our future.
ID: 1374371 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1374378 - Posted: 31 May 2013, 16:03:06 UTC - in response to Message 1374335.  

There doesn't seem to any stock linux app for nvidia, will there be an opti app on Monday?


Will Lunatics_x41g_linux64_cuda32 work with v7?

Frankly we haven't the faintest idea.
It has AC code but it's not been alphaed on Linux.
update your app_info with the tags that have been published and give it a try?

Pay close attention to any problems.

x41g is going to have a higher rate of inconclusives, but it may just work.

One WU finished waiting for validation.

http://setiathome.berkeley.edu/result.php?resultid=3023153356

Will keep an eye one things.
ID: 1374378 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1374379 - Posted: 31 May 2013, 16:09:10 UTC - in response to Message 1374378.  

One WU finished waiting for validation.

http://setiathome.berkeley.edu/result.php?resultid=3023153356

Will keep an eye one things.


Much appreciated here :) Dr Ivan mentioned, buried amidst the chaos, that he's going to have a look for x41zc update needs, and Urs provided some helpful starting suggextions.
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1374379 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1374553 - Posted: 31 May 2013, 22:03:21 UTC - in response to Message 1374210.  
Last modified: 31 May 2013, 22:03:51 UTC

Now that my GTX460 cruncher is on stock app for the first time, it downloaded its first Nvidea AP to crunch using opencl. However, GPU utilization is very low. For MB WUs, GPU utilization is typically 80-90% (for 1 v7 WU), but the AP was only showing ca 4% GPU utilization. I freed up one CPU core and this increased to 18% with occaisional bursts up to 100%. Freeing up a second CPU core made no difference. Is this normal?

The AP WU is currently 24% complete and has been running for 1.5 hours, so it is heading for about 6 hours to complete. Does this sound right for a GTX460 running at 857MHz?

Freeing a core (just one for one card) for OpenCL is normal, yes. Can't comment on timings, I'll leave that for the OpenCL specialists.


Thanks Richard.

In the end it finished normally in 3.5 hours. Is there a clever way of freeing up one CPU core when the GPU is running an AP but when it is running MBs which do not need it? I do not want to loose 25% of my CPU RAC and cannot lways be present to baby sit the machine.

Not sure why the apparent GPU usage was so low, but it has returned to 70-80% now that the GPU is cruching MBs again.


Assuming that was your Task 3022864474, the "percent blanked: 79.34" is why CPU and Run time were long and GPU usage low. More than half the time was spent in class T_remove_radar_randomize which has to be run on CPU.
                                                                  Joe
ID: 1374553 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1374602 - Posted: 1 Jun 2013, 1:35:13 UTC

The developers might want to look at the V7 rollout again with regard to the inability of the servers to supply the necessary files to run the tasks. I'm still pissed that upon updating to V7, the project promptly assigned me 100 CUDA50 tasks while the client was still unsuccessfully trying to download the .DLL for them. And all 100 tasks got errored out immediately because the DLL was still missing. It took over 2 hours to persuade BOINC to download the DLL file. I would have thought the system should not have downloaded any tasks until the necessary files were on board.

Keith

http://setiathome.berkeley.edu/results.php?hostid=5741129&offset=0&show_names=0&state=6&appid=
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1374602 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13886
Credit: 208,696,464
RAC: 304
Australia
Message 1374607 - Posted: 1 Jun 2013, 1:42:32 UTC - in response to Message 1374602.  
Last modified: 1 Jun 2013, 1:43:35 UTC

I would have thought the system should not have downloaded any tasks until the necessary files were on board.

It downloads them, but then holds them untill the necessary files have downloaded. They wouldn't have errored out unless something else happened or was done (such as aborting the download of one or more of the application files).
That's the way it was for me.
Grant
Darwin NT
ID: 1374607 · Report as offensive
Profile betreger Project Donor
Avatar

Send message
Joined: 29 Jun 99
Posts: 11447
Credit: 29,581,041
RAC: 66
United States
Message 1374609 - Posted: 1 Jun 2013, 1:54:37 UTC

Maybe this is the correct thread to ask this question. The auto correlation increases the sensitivity for weak signals by how many db?
ID: 1374609 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1374610 - Posted: 1 Jun 2013, 1:59:33 UTC - in response to Message 1374607.  

I didn't abort the download. I kept hitting try again and watched the backoff times increase, so I shut client down and rebooted. Again watched the client unsuccessfully try to download the .DLL (either unable to access the server or server timeout). Got disgusted and left the room. Came back a couple of hours later and saw the client finally got the .DLL. Only, now the server is only offering up mainly CUDA32 tasks and a couple of CUDA42 tasks. My quota for CUDA50 is now 1 a day until the errors clear.

Cheers, Keith

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1374610 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13886
Credit: 208,696,464
RAC: 304
Australia
Message 1374612 - Posted: 1 Jun 2013, 2:09:51 UTC - in response to Message 1374610.  

I didn't abort the download. I kept hitting try again and watched the backoff times increase, so I shut client down and rebooted. Again watched the client unsuccessfully try to download the .DLL (either unable to access the server or server timeout). Got disgusted and left the room. Came back a couple of hours later and saw the client finally got the .DLL.

Only thing that comes to mind is in it's many attempts to download, one of the files was corrupted or only partial. When it tried to run, it errored out the WUs & re-downloaded the file.

I had something similar happen on one of my systems; on the other system i just left it alone & the files eventually download OK & things started running without aborts or other issues.

Grant
Darwin NT
ID: 1374612 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1374615 - Posted: 1 Jun 2013, 2:44:26 UTC - in response to Message 1374609.  

Maybe this is the correct thread to ask this question. The auto correlation increases the sensitivity for weak signals by how many db?

None. Autocorrelation is a search for a different type of signal, where the transmission is the combination of an original signal combined with a delayed copy of the same signal. Such a combination would be a good indication of intelligent intent given a significant delay.
                                                                   Joe
ID: 1374615 · Report as offensive
juan BFP Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 9786
Credit: 572,710,851
RAC: 3,799
Panama
Message 1374616 - Posted: 1 Jun 2013, 2:47:04 UTC

After 2 days of work, still having some trouble with the scheduler dsending the wrong WU to some hosts, for example cuda50 to a 590/580 host, cuda32 to a 2x690 host, etc. Any clues how to fix that?
ID: 1374616 · Report as offensive
Profile Gary Charpentier Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 25 Dec 00
Posts: 31126
Credit: 53,134,872
RAC: 32
United States
Message 1374617 - Posted: 1 Jun 2013, 2:54:57 UTC - in response to Message 1374615.  

Maybe this is the correct thread to ask this question. The auto correlation increases the sensitivity for weak signals by how many db?

None. Autocorrelation is a search for a different type of signal, where the transmission is the combination of an original signal combined with a delayed copy of the same signal. Such a combination would be a good indication of intelligent intent given a significant delay.
                                                                   Joe

Or gravitational lensing.

ID: 1374617 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 19495
Credit: 40,757,560
RAC: 67
United Kingdom
Message 1374618 - Posted: 1 Jun 2013, 2:59:02 UTC - in response to Message 1374616.  
Last modified: 1 Jun 2013, 3:04:36 UTC

After 2 days of work, still having some trouble with the scheduler dsending the wrong WU to some hosts, for example cuda50 to a 590/580 host, cuda32 to a 2x690 host, etc. Any clues how to fix that?

I'm stuck in the same boat. My two computers, both with 670's, started on 100+ tasks using 32 on one and 42 on the other.

Neither computer has any 50's validated so the server cannot make a decision yet on which is the best app. Although one has got 32 and 42 tasks validated and it should see that the 42 app is much better than 32.

When I get 50's tasks I force the computer to do these asap, in the hope that the more done the greater the chance one, at least, will be verified.
ID: 1374618 · Report as offensive
juan BFP Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 9786
Credit: 572,710,851
RAC: 3,799
Panama
Message 1374619 - Posted: 1 Jun 2013, 3:05:00 UTC - in response to Message 1374618.  

After 2 days of work, still having some trouble with the scheduler dsending the wrong WU to some hosts, for example cuda50 to a 590/580 host, cuda32 to a 2x690 host, etc. Any clues how to fix that?

I'm stuck in the same boat. My two computers, both with 670's, started on 100+ tasks using 32 on one and 42 on the other.

Neither computer has any 50's validated so the server cannot make a decision yet on which is the best app. Although one has got 32 and 42 tasks validated and it should see that the 42 app is much better than 32.

The wierd thing is i allready have a lot of cuda42 work done in the 590/580 host and the scheduler continues to send only cuda50 to that host, the same happens on the 2x690, it allready have a lot of cuda50 done and the server still sending cuda32 to it. Something similar happening of some of my others hosts.

I just don´t understand why not put a instruction to allow us to choose the best app to our host manualy, that could save a lot of time and resources.

ID: 1374619 · Report as offensive
Previous · 1 . . . 11 · 12 · 13 · 14 · 15 · 16 · 17 · Next

Message boards : Number crunching : Now that v7 has rolled out..........


 
©2025 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.