Seti@Home optimized science apps and information
Optimized Seti@Home apps => Linux => Topic started by: Kunin on 15 Jul 2009, 08:45:18 pm
-
Is there one being worked on? Now that I have CUDA working, need something to keep the CPUs busy and haven't gotten any AP v5 yet... thanks.
-
I think they are still waiting for the Linux app to come out of SETI beta first before they release one here.
-
There is one brewing on backstage, but it needs more work.
-
I noticed that the linux app for seti astropulse is now out and its v5.06
-
I noticed that the linux app for seti astropulse is now out and its v5.06
Where you seeing that?
-
Probably here:
Applications (http://setiathome.berkeley.edu/apps.php)
Claggy
-
Probably here:
Applications (http://setiathome.berkeley.edu/apps.php)
Claggy
Ah. I've never been lucky enough to get an AP units.
-
Probably here:
Applications (http://setiathome.berkeley.edu/apps.php)
Claggy
Ah. I've never been lucky enough to get an AP units.
To get them you need to deselect 'Setiathome enhanced' in your preferences And select 'If no work for selected applications is available, accept work from other applications?' to yes,
All because there isn't a APv505 switch in place yet, you'll still get normal enhanced work, and every now and again you'll get an Astropulse v505 wu,
subject to you running Stock apps, or adding the stock AP v505 app (version 506 for Linux) to an app_info.
Claggy
-
To get them you need to deselect 'Setiathome enhanced' in your preferences And select 'If no work for selected applications is available, accept work from other applications?' to yes,
Why only selecting 'If no work for selected applications is available, accept work from other applications?' isn't enough? Whether setiathome enhanced is selected or not shouldn't matter.
-
To get them you need to deselect 'Setiathome enhanced' in your preferences And select 'If no work for selected applications is available, accept work from other applications?' to yes,
Why only selecting 'If no work for selected applications is available, accept work from other applications?' isn't enough? Whether setiathome enhanced is selected or not shouldn't matter.
In which case you'll only get AP_v505 WU's if there's NO setiathome enhanced work available.
In my case, since there's no astropulse or astropulse_v5 availailable, i get issued seti enhanced wu's and sometimes a AP v505 WU.
Claggy
-
Oh yes, I see the difference now. Thanks.
-
There is NOW a AP5.06 version for 64bit linux in public beta testing here at lunatics availbale.
Please don't hesitate to crunch a few wus at SaH/AP beta, this would help greatly with upcoming public release at main.
Thanks.
-
Where about's is it ? I can't find it anywhere !
There is NOW a AP5.06 version for 64bit linux in public beta testing here at lunatics availbale.
Please don't hesitate to crunch a few wus at SaH/AP beta, this would help greatly with upcoming public release at main.
Thanks.
-
Where about's is it ? I can't find it anywhere !
There is NOW a AP5.06 version for 64bit linux in public beta testing here at lunatics availbale.
Please don't hesitate to crunch a few wus at SaH/AP beta, this would help greatly with upcoming public release at main.
Thanks.
It's in beta area.
Sorry, beta testing restricted to only active members of this board.
That is, status "Squire" and higher required AFAIK.
It will available for all soon I hope.
-
I can see the beta section, but all that appears there says AK in it (Win and Refresh), which one is the v5.05?
-
Not in beta downloads, but in the beta section of the forums.
http://lunatics.kwsn.net/18-astropulse-testing/beta-testing-of-astropulse-5-06-for-64bit-linux-with-sse3-capable-cpu.msg20439.html;topicseen#msg20439
Still have to be a squire to access that part of the forum.
-
*sigh*
The lowly live of a knave ...
-
*sigh*
The lowly live of a knave ...
You mean Stormtrooper? :P
-
SNAP !
Without appearing as a spammer trying to get his post count up ( ;) ), how many posts until I get promoted ?
You mean Stormtrooper? :P
-
Without appearing as a spammer trying to get his post count up ( ;) ), how many posts until I get promoted ?
Five, you have two more to go.
Guys, please test the released astropulse clients. If there is no interest in them then there is no point releasing future ones.
-
I would love too, but I do not have any machines currently running Linux at the moment.
-
Without appearing as a spammer trying to get his post count up ( ;) ), how many posts until I get promoted ?
Five, you have two more to go.
Guys, please test the released astropulse clients. If there is no interest in them then there is no point releasing future ones.
I just d/l'd the AMD one and will be setting it up right now to test it on my dual core Ath64...
-
Well I'm willing to test. I'm gonna up my posts, then download when I get home from work tonight.
-
Last one (I hope there's AP units available !)
-
Downloaded and installed.
I'm assuming this version isn't compatible with 505 units ?
Is there anything I need to do to my boinc/seti config to get 506 WUs ?
-
Don't be fooled by the name, it is really a 5.05 client.
You have to change seti@home preferences in your account so that "If no work for selected applications is available, accept work from other applications". Also turn off seti@home enhanced in the selected applications list.
-
Don't be fooled by the name, it is really a 5.05 client.
You have to change seti@home preferences in your account so that "If no work for selected applications is available, accept work from other applications". Also turn off seti@home enhanced in the selected applications list.
Since Urs has posted:
NOT FOR RELEASE. Not Intended for distribution or use on main.
and the Seti Beta project only uses Astropulse, and Doesn't use Astropulse_v5 or Astropulse_v505,
you Don't need to set "If no work for selected applications is available, accept work from other applications" to yes, leave it at no,
just make sure you select Astropulse to yes (the Astropulse_5 entry doesn't do anything) and deselect setiathome enhanced.
Claggy
Edit: see: optimized AP (http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=1575&nowrap=true#37790)
-
Don't be fooled by the name, it is really a 5.05 client.
You have to change seti@home preferences in your account so that "If no work for selected applications is available, accept work from other applications". Also turn off seti@home enhanced in the selected applications list.
It might be irritating a bit that linux has a different application name, but that is also the case with the standard application. Check the SaH applications page (http://setiathome.berkeley.edu/apps.php) for the actual version naming. Important is only that 5.05/5.06 produce comparable results for the validation process.
-
So I've been running AP overnight. 4 units completed, none validated, but 1 did fail with a computation error :
http://setiathome.berkeley.edu/result.php?resultid=1339606984
http://setiathome.berkeley.edu/result.php?resultid=1339613053
Hopefully there's some info in there for debug. My CPU isn't overclocked, and isn't overheating, so it's probably not that. It may be a completely random error though ...
*EDIT*
2 computarion errors.
Is this the best place to report these errors ?
-
I don't think you can run it on the main board yet, it is only for the Beta at the moment.
-
I see you're using kernel 2.6.28-15-generic, this is the ubuntu 9.04 kernel right? For some reason, this kernel isn't good for running lunatics's optimised apps. Revert back to 8.10's kernel (2.6.27-14) or build your own.
-
Yeah, it's ubuntu.
Don't have any issues with other apps (the CUDA apps crunch fine ...).
Surely a kernel I build myself will be pretty much identical though ?
-
Surely a kernel I build myself will be pretty much identical though ?
Not necessarily.
I don't know if this bug exists in ubuntu's compiled kernel or 2.6.28 in general. Other people with the same problem, compiling their own newer kernel (2.6.29, 2.6.30) had this problem go away. I haven't bothered with compiling my own kernel, I'm running 9.04 with 8.10's 2.6.27-14 kernel.
-
I don't think you can run it on the main board yet, it is only for the Beta at the moment.
Yes, build and packaged (app_info.xml) for Beta.
-
So I've been running AP overnight. 4 units completed, none validated, but 1 did fail with a computation error :
http://setiathome.berkeley.edu/result.php?resultid=1339606984
http://setiathome.berkeley.edu/result.php?resultid=1339613053
Hopefully there's some info in there for debug. My CPU isn't overclocked, and isn't overheating, so it's probably not that. It may be a completely random error though ...
*EDIT*
2 computarion errors.
Is this the best place to report these errors ?
Many thanks for beta testing this application version.
The best and preferred place for reports is the forum thread where you find the AP apps.
These error looks pretty similar to that one which was seen with previous (v5.03) versions and exactly this 2.6.28 kernel. Hopefully the trace contains enough hints to find out why it fails.
If you don't mind, could you try the other application version as well ?
-
I'll try that tomorrow.
At the moment I'm trying the SSE3 version on a 2.6.30 kernel, just wanna make sure that works.
... by other version, you mean the SSE2 only version, correct ? Which kernel ? The 2.6.28 one ?
-
I'll try that tomorrow.
At the moment I'm trying the SSE3 version on a 2.6.30 kernel, just wanna make sure that works.
... by other version, you mean the SSE2 only version, correct ? Which kernel ? The 2.6.28 one ?
If 2.6.30 works for you without errors, that would at least proove that 2.6.28 has to be excluded from support, similar to kernel versions before 2.6.22.
Yes, the AMD64 SSE2 version. If possible, run a few wus on each kernel, to see if there will occur other problems. Thanks in advance.
ps: your first wu has validated. ++
-
I'll keep posting here, as this seems the more active thread :)
The vanilla 2.6.28 kernel also resulted in errors, so I don't think its an ubuntu thing.
The SSE2 client seemed to work fine, but when I switched back to SSE3, the executable wasn't in the folder, and I lost my cache of AP units !!! Bugger !
I'm going to try the 2.6.31rc6 kernel, with the new nvidia drivers, see if I can hit the middle ground of stable AP, and fast CUDA.
-
I'll keep posting here, as this seems the more active thread :)
The vanilla 2.6.28 kernel also resulted in errors, so I don't think its an ubuntu thing.
The SSE2 client seemed to work fine, but when I switched back to SSE3, the executable wasn't in the folder, and I lost my cache of AP units !!! Bugger !
I'm going to try the 2.6.31rc6 kernel, with the new nvidia drivers, see if I can hit the middle ground of stable AP, and fast CUDA.
The only reported errors after one week into beta testing are yours (Ubuntu 9.04, kernel version 2.6.28+), lordvader. A big thank you for your efforts.
Anyone else has results to report ?
ps: stock astropulse seems not to play well on that os/kernel (http://setiathome.berkeley.edu/forum_thread.php?id=55154), too.
-
On the AMD client, I get mostly good runs, one error. But, they all had very low claimed credit:
Task ID
click for details
Show names Work unit ID
click for details Sent Time reported
or deadline
explain Status Run time (sec) claimed credit granted credit
6627643 2282517 19 Aug 2009 23:36:57 UTC 20 Aug 2009 0:08:22 UTC Completed, validation inconclusive 1,286.88 3.13 pending
6627635 2282513 19 Aug 2009 23:28:59 UTC 20 Aug 2009 0:02:44 UTC Completed, waiting for validation 396.70 0.97 pending
6627619 2282505 19 Aug 2009 23:13:26 UTC 19 Aug 2009 23:28:59 UTC Completed, validation inconclusive 161.91 0.38 pending
6627609 2282500 20 Aug 2009 0:08:22 UTC 14 Sep 2009 0:08:22 UTC In progress --- --- ---
6627595 2282493 20 Aug 2009 0:02:44 UTC 20 Aug 2009 9:14:29 UTC Completed, waiting for validation 291.70 0.68 pending
6627490 2282443 19 Aug 2009 23:16:42 UTC 19 Aug 2009 23:36:57 UTC Completed, validation inconclusive 636.87 1.55 pending
6624580 2281948 17 Aug 2009 21:22:49 UTC 18 Aug 2009 15:26:25 UTC Error while computing 2,295.81 8.93 8.93
-
On the AMD client, I get mostly good runs, one error. But, they all had very low claimed credit:
Task ID
click for details
Show names Work unit ID
click for details Sent Time reported
or deadline
explain Status Run time (sec) claimed credit granted credit
6627643 2282517 19 Aug 2009 23:36:57 UTC 20 Aug 2009 0:08:22 UTC Completed, validation inconclusive 1,286.88 3.13 pending
6627635 2282513 19 Aug 2009 23:28:59 UTC 20 Aug 2009 0:02:44 UTC Completed, waiting for validation 396.70 0.97 pending
6627619 2282505 19 Aug 2009 23:13:26 UTC 19 Aug 2009 23:28:59 UTC Completed, validation inconclusive 161.91 0.38 pending
6627609 2282500 20 Aug 2009 0:08:22 UTC 14 Sep 2009 0:08:22 UTC In progress --- --- ---
6627595 2282493 20 Aug 2009 0:02:44 UTC 20 Aug 2009 9:14:29 UTC Completed, waiting for validation 291.70 0.68 pending
6627490 2282443 19 Aug 2009 23:16:42 UTC 19 Aug 2009 23:36:57 UTC Completed, validation inconclusive 636.87 1.55 pending
6624580 2281948 17 Aug 2009 21:22:49 UTC 18 Aug 2009 15:26:25 UTC Error while computing 2,295.81 8.93 8.93
[/size]
Are you sure you had the optimized version running on that error (http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=6624580). I have checked the task id :<core_client_version>6.6.11</core_client_version>
<![CDATA[
<message>
process got signal 11
</message>
<stderr_txt>
In ap_gfx_main.cpp: in ap_graphics_init(): Starting client.
In ap_client_main.cpp: in mainloop(): at dm_chunk_large 896
In ap_client_main.cpp: in mainloop(): at dm_chunk_large 1024
SIGSEGV: segmentation violation
</stderr_txt>
]]>This is the stock astropulse 5.06 application having an error not the optimized one. All of your results, Tye, look like that.
If you have problems with the installation instructions that were included in the optimized package, please ask for help in this forum. There are a few people here that can assist you through the installation process.
The low claimed credts are due to very low runtimes because of early finishing with "Found 30 single pulses and 30 repeating pulses, exiting."
-
You're right - that was 2 days before I switched. The others should be with the new client, or at least the ones 20 Aug and later. It shows up in top and ps with the correct client and has since I installed it. Hmmm. I'll check more after work tonight.
-
Aha, it turned out I hadn't completed any yet. Just submitted one right now manually and it looks good:
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=6627609
-
Aha, it turned out I hadn't completed any yet. Just submitted one right now manually and it looks good:
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=6627609
Thanks Tye, that one looks indeed like the right thing. :)
-
I was wondering if the AP issues exist under kernel 2.6.27 ?
Moving to kernel 2.6.30 won't cut it, as CUDA performance in abysmal, and the AP failure rate in 2.6.28 is just horrible.
I haven't tried 2.6.29 though ...
-
lordvader and others,
to find out about the reported errors i did setup and now use this host (http://setiweb.ssl.berkeley.edu/beta/results.php?hostid=41956) at seti beta using Ubuntu 9.04 with kernel 2.6.28 .
If the error does occur within the next 15 to 20 wus i hope i can identify the problem exactly. Otherwise i have to put a warning for Ubuntu users in the package and release to public.
There is parallel running an identical pc using openSuSE11.1 with kernel 2.6.27 which has done a few wus already without any problems.
Let's see what comes. ;D
-
Well I'm currently running 2.6.27-31.
Have 2 valid units, 1 pending (inconclusive), and 4 running.
None have bailed out with error, so far so good.
Kinda surprised that your host hasn't errored yet. is it possible that the issue is a combo of 2.6.28 and amd64 cpus ?
... the output of your tasks look different to mine ... are you running the intel only version ?
(i'm not getting the "In ap_client_main.cpp: in mainloop(): at dm_chunk_large" strings)
-
Well I'm currently running 2.6.27-31.
Have 2 valid units, 1 pending (inconclusive), and 4 running.
None have bailed out with error, so far so good.
Kinda surprised that your host hasn't errored yet. is it possible that the issue is a combo of 2.6.28 and amd64 cpus ?
... the output of your tasks look different to mine ... are you running the intel only version ?
(i'm not getting the "In ap_client_main.cpp: in mainloop(): at dm_chunk_large" strings)
That is a debug version, so if the error occurs i hope to get more exact info where. Sorry for confusing you.
:) Very nice that 2.6.27 is not erroring out so far (that is what i expected). Must really be a problem with Ubuntu 9.04 and newer kernel version from 2.6.28 up. Tye's amd64 also did not throw an error, but he has done only one wu as of yet.
Correction: Tye has finished a second wu with 2.6.28 with no error and now uses a newer kernel 2.6.29.
Update: Three 32bit versions are about to enter the beta phase tomorrow. See other thread then.
Update2: Stopped using debug version, because the errors did not appear to my host.
-
Hy Urs,
any need for additional tests for the linux 32 bit ?
-
2nd post ;)
Matthias
-
3rd post ;)
can test under ubuntu 8.10
Matthias
-
4st post ::)
if needed I also have a debian 32 bit system
Matthias
-
4st post ::)
if needed I also have a debian 32 bit system
Matthias
Thanks Matthias, but you are a little late. ;) Sorry, no additional tests needed. Release will come in the next days.
Only problem that showed up was with Ubuntu9.04 and derivatives and kernels 2.6.28 and up.
That was known before beta testing started. It is now clear that this is not our fault, thanks to lordvader running some extra tests.
-
4st post ::)
if needed I also have a debian 32 bit system
Matthias
Thanks Matthias, but you are a little late. ;) Sorry, no additional tests needed. Release will come in the next days.
Only problem that showed up was with Ubuntu9.04 and derivatives and kernels 2.6.28 and up.
That was known before beta testing started. It is now clear that this is not our fault, thanks to lordvader running some extra tests.
Hi Urs, thanks for the reply.
nice to hear that the tests are done.
Linux 64 bit I've activated in my 64 bit machines. :D
Matthias
-
Hi Urs, thanks for the reply.
nice to hear that the tests are done.
Linux 64 bit I've activated in my 64 bit machines. :D
Matthias
Don't forget to check in a few days that you still have the most actual version. ;)
-
Aren't these new APs giving less credit? I got 1200 for the older v5 APs but only 800 for the new ones. They take 13 hours so ~60 credit per hour. Crunching VLARs gives me 50-100 credit per hour depending on the angle range (0.01-0.13). Strangely enough I crunched a few APs a week ago that gave 799 credits but only took 10 hours and that's why I tried to focus on APs instead of VLARs but I'll probably switch back to VLARs now...
-
Aren't these new APs giving less credit? I got 1200 for the older v5 APs but only 800 for the new ones. They take 13 hours so ~60 credit per hour. Crunching VLARs gives me 50-100 credit per hour depending on the angle range (0.01-0.13). Strangely enough I crunched a few APs a week ago that gave 799 credits but only took 10 hours and that's why I tried to focus on APs instead of VLARs but I'll probably switch back to VLARs now...
Yes, erm, 'some jokers' went and made stock go faster, which raises the performance of the 'median machine', bringing the credits down. Don't worry, further improvements to the optimised codebase are relatively esoteric, and so won't 'readily' translate to stock, though both variants are liekly to get some fundamental overhauls over time, optimised will pull gradually further ahead again.
-
So you recommend I stay with AP for now or would it be more useful to the project if I crunched the VLARs again for a while?
-
Hmm, well looking at the current Arecibo situation, it might pay to squirrel away whatever you can get But Personally, on my machines, Astropulse seems to pay about equal per time unit to multibeam, therefore it comes down to what you prefer, versus available computation capacity.
One possible way of looking at it is that in the past AstoPulse has only been far higher paying (with optimised) due to lacklustre performance of he stock app. Now that it is somewhat 'cleaned up', stock is partially optimised, therefore optimised is closer to stock.
Possibly Limited remaining work means though, that choosing one app over another turns off one of two 'taps', so limiting to one kind would see you short of work sooner then if you'd just processed both kinds.
Of course I'm trying to look at it from a productivity point of view, rather than a credit one. If I was looking at credits I;d probably be doing Milky way or something, but for everyday use in daily life I figure that my mental picture of a spiral thing made out of dots is good enough of a model...
-
The problem I have (and the rest of you too I guess) is that if I add AK_V8 to my app_info again, to be able to crunch multibeam on CPU, BOINC will also start downloading MB to CPU queue. When I then reschedule the VLARs from my CUDA-queue, BOINC goes into EDF mode and won't download neither AP nor MB to CPU. Getting APs after that is extremely difficult since BOINC seems to prefer downloading MB to CPU whenever there is deficiency. I wish there was a way to tell BOINC not to download MB specifically to CPU but still allow them to be crunched... but I guess I'll go back to AK_V8 then instead of risking running out of work. Thanks for your valuable input Jason.
-
Sorry to interrupt, but I am working with 6.6.38 and there has been noted that Debts (Long and Short Term) are having issues. So while the machine I am monitoring is running Seti Beta. I am seeing the effects. So as I can not reschedule Seti Beta, I can see where loading the CPU is going to drive The Debts Nuts (which would partially explain the EDF)....