Updating for upcoming apps.....

Message boards : Number crunching : Updating for upcoming apps.....
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 5 · 6 · 7 · 8 · 9 · 10 · Next

AuthorMessage
Profile arkayn
Volunteer tester
Avatar

Send message
Joined: 14 May 99
Posts: 4438
Credit: 55,006,323
RAC: 0
United States
Message 1208306 - Posted: 21 Mar 2012, 0:49:30 UTC

I got one on my GTX-460 machine, only because I have the older Nvidia AP app on it.

ID: 1208306 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13720
Credit: 208,696,464
RAC: 304
Australia
Message 1208389 - Posted: 21 Mar 2012, 6:39:13 UTC - in response to Message 1208198.  

Why does the installer not insert <flops> lines?

Because, eventually, they'll fix the DCF problem. Then there'll be all sorts of hell involved with WUs timing out etc.
Grant
Darwin NT
ID: 1208389 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1208614 - Posted: 21 Mar 2012, 22:58:46 UTC - in response to Message 1208607.  

if the installer has been delayed cause Eric ... why not take time to put Nvidia Cuda AP6 appl in it ? /p

:D

Because it's still in Alpha testing and isn't good enough to be put in the installer yet, when the 100% CPU Bug has been fixed and it's fully thread safe then it'll be in the installler, and not a moment before that.

Claggy
ID: 1208614 · Report as offensive
Oddbjornik Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 15 May 99
Posts: 220
Credit: 349,610,548
RAC: 1,728
Norway
Message 1208619 - Posted: 21 Mar 2012, 23:03:44 UTC

It looks like SETI V7 is out. Splitters are running, cricket is maxed, and at the same time "old SETI" workunits out in the field on the stats page is slowly sinking.

I guess this means that everyone running the Lunatics apps should prepare to panic! We won't be getting much new work until we get some new apps!
ID: 1208619 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1208621 - Posted: 21 Mar 2012, 23:08:25 UTC - in response to Message 1208619.  
Last modified: 21 Mar 2012, 23:09:21 UTC

It looks like SETI V7 is out. Splitters are running, cricket is maxed, and at the same time "old SETI" workunits out in the field on the stats page is slowly sinking.

I guess this means that everyone running the Lunatics apps should prepare to panic! We won't be getting much new work until we get some new apps!

Unlikely since there is no v7 apps listed on the applications page, more likely a combination of Astropulse v6 and a shortie storm causing the cricket to be maxed out and the status page to show minimal MB tasks ready to send.

Claggy
ID: 1208621 · Report as offensive
Oddbjornik Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 15 May 99
Posts: 220
Credit: 349,610,548
RAC: 1,728
Norway
Message 1208626 - Posted: 21 Mar 2012, 23:19:45 UTC - in response to Message 1208621.  

Hmm... I'm not seeing any shortie storm on my computers, and the result creation rate is low...

Nothing to do but wait and see, I think.
ID: 1208626 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1208755 - Posted: 22 Mar 2012, 7:50:45 UTC

Well, a nice "old apps" comparison with comparable hardware.
This guy running an i7-2600K, default project application, 94,870 seconds.
Me running an i5-2500K, Lunatics ap_5.05r409_SSE application, 24,542 seconds.

Chapeau.
ID: 1208755 · Report as offensive
LadyL
Volunteer tester
Avatar

Send message
Joined: 14 Sep 11
Posts: 1679
Credit: 5,230,097
RAC: 0
Message 1208759 - Posted: 22 Mar 2012, 8:16:28 UTC - in response to Message 1208755.  

Well, a nice "old apps" comparison with comparable hardware.
This guy running an i7-2600K, default project application, 94,870 seconds.
Me running an i5-2500K, Lunatics ap_5.05r409_SSE application, 24,542 seconds.

Chapeau.


You can expect a few more % from AP v6.01 r555/r557 and some 20-25% extra if you have AVX ;)
I'm not the Pope. I don't speak Ex Cathedra!
ID: 1208759 · Report as offensive
Profile cliff
Avatar

Send message
Joined: 16 Dec 07
Posts: 625
Credit: 3,590,440
RAC: 0
United Kingdom
Message 1209380 - Posted: 23 Mar 2012, 20:47:35 UTC

Can anyone tell me what happens if an app_info.xml get removed/corrupted without it being uninstalled or the project reset?

Will Boinc simply revert to using stock apps if it cant find the app_info file or will it go bonkers and error out any WU in the cache?

Cheers,
Cliff,
Been there, Done that, Still no damm T shirt!
ID: 1209380 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 1209397 - Posted: 23 Mar 2012, 21:35:16 UTC
Last modified: 23 Mar 2012, 21:37:52 UTC

I tried that at first. Just simply deleting it doesn't really appear to do a whole lot. Deleting the EXEs that were described in the app_info really annoys BOINC and it then tries to re-download them.

What a project reset does is goes into client_state and basically purges everything except host-specific stuff (computer ID, account, etc).

I cleared it out manually by going and finding the two or three entries in client_state that referenced r409 and just simply removed them. Once I got assigned a new v6 AP, all the needed files were downloaded automatically.

On hindsight, since I did all of that with an empty cache, hitting the reset button would have been really easy. I believe it gets slightly (and I do mean ever-so-slightly) trickier when going stock->opt. Do the app_info and put the new EXEs in the right place, and then delete the stock ones, and it's -supposed- to figure it out and keep going.

edit: Realized I didn't exactly answer your question. As I was told earlier in this thread, if you just simply hit the reset button, tasks that were partially-crunched and ones waiting to run will be purged. You -should- get them re-sent to you, along with the stock apps.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 1209397 · Report as offensive
Profile cliff
Avatar

Send message
Joined: 16 Dec 07
Posts: 625
Credit: 3,590,440
RAC: 0
United Kingdom
Message 1209402 - Posted: 23 Mar 2012, 21:51:20 UTC - in response to Message 1209397.  

Right:-) What concerned me was that Vince got hammered with a nasty the other day, and it occured to me that if some rootkit or other nasty gets onto a rig who knows what it screws up, moves or simply deletes.

And I do a fair bit of sniffing around the web..

Methinks a backup protocol for boinc is required.. even if its just the app_info and exes, and an archive of the sub dir that the lunatics installer puts the stock apps in. Just in case.

Regards,
Cliff,
Been there, Done that, Still no damm T shirt!
ID: 1209402 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1209405 - Posted: 23 Mar 2012, 22:01:43 UTC - in response to Message 1209402.  
Last modified: 23 Mar 2012, 22:02:03 UTC

Most rootkits won't give a damn about your BOINC data directory. Not unless someone with a BOINC hatred writes one that specifically attacks BOINC data directories, do I think you need to worry about anything. If you want to backup BOINC, backup the full data directory, with all sub-directories and files.

Else, BOINC has built in redundancy where if you can't return the work, it'll time out and be sent to another computer which will do the work instead.

And if your system does get compromised in some form or way, after you cleaned up the mess, you can usually just reinstall BOINC and continue where you left off (due to the work being in the separate data directory). Having the data directory live on a disk or partition separate from your Windows partition is without doubt the best solution. Then only a full drive fail will break your ability to return the work.
ID: 1209405 · Report as offensive
Profile Cliff Harding
Volunteer tester
Avatar

Send message
Joined: 18 Aug 99
Posts: 1432
Credit: 110,967,840
RAC: 67
United States
Message 1209410 - Posted: 23 Mar 2012, 22:14:18 UTC - in response to Message 1209402.  

Right:-) What concerned me was that Vince got hammered with a nasty the other day, and it occured to me that if some rootkit or other nasty gets onto a rig who knows what it screws up, moves or simply deletes.

And I do a fair bit of sniffing around the web..

Methinks a backup protocol for boinc is required.. even if its just the app_info and exes, and an archive of the sub dir that the lunatics installer puts the stock apps in. Just in case.

Regards,


I do something like this during my nightly backup schedule. My data directory resides an a separate physical drive from the EXEs, which are on the system drive. The backup includes an image of the system drive and all of the data drive except the projects, slots, & symbols folders.

By the way does anybody know if David has had a chance to look at the GPU zero results problem?


I don't buy computers, I build them!!
ID: 1209410 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1209411 - Posted: 23 Mar 2012, 22:16:35 UTC - in response to Message 1209397.  

I believe it gets slightly (and I do mean ever-so-slightly) trickier when going stock->opt. Do the app_info and put the new EXEs in the right place, and then delete the stock ones, and it's -supposed- to figure it out and keep going.

Even easier than that. Place the app_info. Place the new exes. Restart BOINC.

No need to delete anything. On Windows, you could even get the Lunatics Installer to do all three steps for you, including - from next week - the apps for Astropulse v6..
ID: 1209411 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1209413 - Posted: 23 Mar 2012, 22:22:04 UTC - in response to Message 1209410.  

By the way does anybody know if David has had a chance to look at the GPU zero results problem?

Do you mean the 'zero credit for results' problem that was holding us up?

I'm just looking at changeset [trac]changeset:25484[/trac], posted half an hour ago:

scheduler: fix bug that could cause zero credit for the first few jobs of a new application
(in wu_estimated_pfc(), only multiply by app.min_avg_pfc if it's nonzero).

That sounds like the one, with a plausible resolution.

ID: 1209413 · Report as offensive
Profile Cliff Harding
Volunteer tester
Avatar

Send message
Joined: 18 Aug 99
Posts: 1432
Credit: 110,967,840
RAC: 67
United States
Message 1209458 - Posted: 24 Mar 2012, 1:26:08 UTC - in response to Message 1209413.  

By the way does anybody know if David has had a chance to look at the GPU zero results problem?

Do you mean the 'zero credit for results' problem that was holding us up?

I'm just looking at changeset [trac]changeset:25484[/trac], posted half an hour ago:

scheduler: fix bug that could cause zero credit for the first few jobs of a new application
(in wu_estimated_pfc(), only multiply by app.min_avg_pfc if it's nonzero).

That sounds like the one, with a plausible resolution.



That's very good news. Tell David THANKS!! from all of us.


I don't buy computers, I build them!!
ID: 1209458 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1209627 - Posted: 24 Mar 2012, 13:15:46 UTC

Hmmm, now, not impressed. Had two of those AP v6.01s and they both ran for all of 2 seconds. What's the use of downloading 16MB worth of data that runs all together 4 seconds, less even when you count CPU time (0.39 + 0.41 seconds)? ;-)
ID: 1209627 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1209629 - Posted: 24 Mar 2012, 13:21:43 UTC - in response to Message 1209627.  
Last modified: 24 Mar 2012, 13:24:23 UTC

Hmmm, now, not impressed. Had two of those AP v6.01s and they both ran for all of 2 seconds. What's the use of downloading 16MB worth of data that runs all together 4 seconds, less even when you count CPU time (0.39 + 0.41 seconds)? ;-)

Matt's Technical News post from Thursday:

Other progress: we finally launched the new splitters which are sensitive to VGC values and thus skip (most) noisy data blocks instead of splitting them into workunits that will return quickly and clog up our pipeline. Yay! However there were unexpected results last night: turns out it's actually slower to parse such a noisy data file and skip bad blocks than to just split everything, so splitters were getting stuck on these files and not generating work. Oops. We ran out of multi-beam work last night due to this, and I reverted back this morning just to the plumbing working again. I'm going to change the logic to be a little more aggressive and thus speed up skipping through noisy files, and implement that next week.

At least it won't bump your APR up to unbelievable high values now,

Claggy
ID: 1209629 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1209638 - Posted: 24 Mar 2012, 13:31:42 UTC - in response to Message 1209629.  

At least it won't bump your APR up to unbelievable high values now,

No, they left me with no APR. What is the unbelievably high value, though?
ID: 1209638 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1209645 - Posted: 24 Mar 2012, 13:49:19 UTC - in response to Message 1209638.  

At least it won't bump your APR up to unbelievable high values now,

No, they left me with no APR. What is the unbelievably high value, though?

I think we've seen it up into the TeraFlop range. 2 TFlop, IIRC.
ID: 1209645 · Report as offensive
Previous · 1 . . . 5 · 6 · 7 · 8 · 9 · 10 · Next

Message boards : Number crunching : Updating for upcoming apps.....


 
©2024 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.