Updating for upcoming apps.....

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

To post messages, you must log in.

Previous · 1 . . . 3 · 4 · 5 · 6 · 7 · 8 · 9 . . . 10 · Next

AuthorMessage
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1206666 - Posted: 16 Mar 2012, 19:10:48 UTC - in response to Message 1206663.  
Last modified: 16 Mar 2012, 19:11:20 UTC

Speaking about AP v6. This workunit caught my attention. Look at _0, he got less that a day to finish.
That's strange, haven't seen it in any of my other AP v6.
http://setiathome.berkeley.edu/workunit.php?wuid=949641584

A well known side effect of the resend lost work mechanism is that an MB .vlar task may be expired immediately when the Scheduler is checking whether it can resend the task to a CUDA GPU, see message 1112328. This is probably the same thing, an AP v6 task also cannot be sent to a CUDA GPU.
                                                                   Joe

Might be. Though I have never seen it on an AP workunit.

At present there are no VLAR AP tasks. Only MB tasks have that distinction.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1206666 · Report as offensive
Profile cliff
Avatar

Send message
Joined: 16 Dec 07
Posts: 625
Credit: 3,590,440
RAC: 0
United Kingdom
Message 1206670 - Posted: 16 Mar 2012, 19:18:59 UTC - in response to Message 1206648.  

Hi Richard,
I wonder could the server perhaps be used to send at leat one warning message to that effect when it 1st detects an app_info.xml file?
And maybe repeat the warning whenever a new stock app is being distributed.

Make more sense than the current and too oft repeated you dont have a valid app in your app_info.xml file:-)

Perhaps it should send, "you are using an app_info.xml file and should check to see if there are updated apps for your system."

Regards,
Cliff,
Been there, Done that, Still no damm T shirt!
ID: 1206670 · 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 1206677 - Posted: 16 Mar 2012, 19:43:21 UTC - in response to Message 1206666.  

Speaking about AP v6. This workunit caught my attention. Look at _0, he got less that a day to finish.
That's strange, haven't seen it in any of my other AP v6.
http://setiathome.berkeley.edu/workunit.php?wuid=949641584

A well known side effect of the resend lost work mechanism is that an MB .vlar task may be expired immediately when the Scheduler is checking whether it can resend the task to a CUDA GPU, see message 1112328. This is probably the same thing, an AP v6 task also cannot be sent to a CUDA GPU.
                                                                   Joe

Might be. Though I have never seen it on an AP workunit.

At present there are no VLAR AP tasks. Only MB tasks have that distinction.

I think Joe was saying:

"In the case of MB, most tasks can be (re-)sent to CUDA. But a small fraction of them - the VLAR group - can't be (re-)sent to CUDA, and are prone to having the deadline manipulated.

"In the case of AP, all tasks can't be sent to CUDA. So all AP tasks are similar to VLAR tasks in that regard."
ID: 1206677 · 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 1206678 - Posted: 16 Mar 2012, 19:45:46 UTC - in response to Message 1206670.  

Hi Richard,
I wonder could the server perhaps be used to send at leat one warning message to that effect when it 1st detects an app_info.xml file?
And maybe repeat the warning whenever a new stock app is being distributed.

Make more sense than the current and too oft repeated you dont have a valid app in your app_info.xml file:-)

Perhaps it should send, "you are using an app_info.xml file and should check to see if there are updated apps for your system."

Regards,

That's a thought. Do you mind if I hang on to it and explore it once this launch is out of the way. I'm not quite sure yet whether it needs to be directed to the SETI or to the BOINC development team for consideration.
ID: 1206678 · Report as offensive
Profile red-ray
Avatar

Send message
Joined: 24 Jun 99
Posts: 308
Credit: 9,029,848
RAC: 0
United Kingdom
Message 1206683 - Posted: 16 Mar 2012, 19:55:42 UTC - in response to Message 1206678.  
Last modified: 16 Mar 2012, 20:03:17 UTC

Hi Richard,

If you are going to fix it then the following is either very annoying or amusing depending on how you look at it.

16/03/2012 19:06:08 | SETI@home | Message from server: Your app_info.xml file doesn't have a usable version of AstroPulse v6.
16/03/2012 19:06:08 | SETI@home | This computer has reached a limit on tasks in progress
ID: 1206683 · Report as offensive
Profile cliff
Avatar

Send message
Joined: 16 Dec 07
Posts: 625
Credit: 3,590,440
RAC: 0
United Kingdom
Message 1206696 - Posted: 16 Mar 2012, 20:47:57 UTC - in response to Message 1206678.  
Last modified: 16 Mar 2012, 20:49:38 UTC

Hi Richard,
By all means, thats pretty much why I posted:-)
A wee bit of my old beta testing mindset is slowly starting to wriggle out of storage:-) My guess would be the Seti team, afterall its their server:-)

But its got a long way to wriggle.. going on a decade and a half:-) So dont expect to many good ideas from this direction:-)

Regards,
Cliff,
Been there, Done that, Still no damm T shirt!
ID: 1206696 · Report as offensive
OTS
Volunteer tester

Send message
Joined: 6 Jan 08
Posts: 369
Credit: 20,533,537
RAC: 0
United States
Message 1206708 - Posted: 16 Mar 2012, 21:38:20 UTC - in response to Message 1206647.  

Richard wrote in his post 1206160 earlier in this thread:

According to the Applications page, both Linux/x86 (32 bit) and Linux/x86_64 are already installed and available as stock apps. You can be the first to try them!


and Joe later wrote in his post 1206647:


Any host which has an app_info.xml is not eligible to get stock applications automatically. Stock applications are made for a specific platform, going anonymous platform means the user must supply the applications and tell BOINC what work they can do. For AP v6 that is not yet a good thing to do, it causes zero credit grants for both the anonymous platform host and the wingmate.


The first quote, if I am reading it right, indicates that perhaps I should try the stock Linux app and see how it works while the second seems to indicate that doing so might not be such a good idea. Which would be the best approach to take?

Thanks to both of you for taking the time to explain as much as you have about how things work and also to Urs for the update on the new optimized Linux AP apps. I appreciate it.
ID: 1206708 · Report as offensive
Profile S@NL Etienne Dokkum
Volunteer tester
Avatar

Send message
Joined: 11 Jun 99
Posts: 212
Credit: 43,822,095
RAC: 0
Netherlands
Message 1206717 - Posted: 16 Mar 2012, 22:09:25 UTC - in response to Message 1206534.  

Speaking about AP v6. This workunit caught my attention. Look at _0, he got less that a day to finish.
That's strange, haven't seen it in any of my other AP v6.
http://setiathome.berkeley.edu/workunit.php?wuid=949641584


been happening to "normal"MB tasks too lately, got 2 vlar's today with less then an hour to report I noticed. Must be a glitch seperate from the roll-out of AP 6.01
ID: 1206717 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 1206722 - Posted: 16 Mar 2012, 22:29:37 UTC - in response to Message 1206708.  
Last modified: 16 Mar 2012, 22:36:30 UTC

The first quote, if I am reading it right, indicates that perhaps I should try the stock Linux app and see how it works while the second seems to indicate that doing so might not be such a good idea. Which would be the best approach to take?

Thanks to both of you for taking the time to explain as much as you have about how things work and also to Urs for the update on the new optimized Linux AP apps. I appreciate it.

If you really want to do some of this v6 work and help the validators establish a baseline sooner, you can certainly go stock. This means removing app_info and resetting the project (there's a button in BOINC for reset). I don't believe you have to have an empty cache to do it, but it would be recommended.

This means you will be running completely stock for both MB (and CUDA), and AP. Once the Lunatics AP app gets released, you can switch back to having an app_info and optimized applications.

As was explained earlier in the thread, you can't run optimized MB and stock AP. Using the stock AP app within an app_info still makes it anonymous platform and will result in crazy credit situations.

The only down-side to using the stock app is that the run time is significantly longer--usually about double. For instance, my machine was doing v505 tasks in about 45,000 seconds, but the stock app for v6 is running right around 100,000. Twice the run-time for the (approximately) same amount of credit = half the credit:time ratio. That only matters if you care about credit. If you mostly only care about valid science, then it shouldn't matter either way. Of course with optimized, you can get twice the science done in the same amount of time. :p



regarding the super-short deadline on that one AP, it was probably a case where it would have been re-sent, but BOINC was only asking for GPU work, so the scheduler nuked it (as it does with VLAR on the MB-side of things).

edit: oh, and I have two more validations. So far with the three that I've got, the credit seems to be pretty consistent in the 540-560 range.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 1206722 · Report as offensive
Profile Jim_S
Avatar

Send message
Joined: 23 Feb 00
Posts: 4705
Credit: 64,560,357
RAC: 31
United States
Message 1207006 - Posted: 17 Mar 2012, 16:10:51 UTC

Please Post Beta Info at the Beta Boards!

I Desire Peace and Justice, Jim Scott (Mod-Ret.)
ID: 1207006 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1207090 - Posted: 17 Mar 2012, 17:45:01 UTC

Geez, I have 9 rigs running......and see Boinc messages on none of them.
So, it does not bother me at all if the severs whine about not having the right app for v6. Have at it.....LOL.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1207090 · 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 1207200 - Posted: 17 Mar 2012, 21:49:12 UTC

Calling Lunatics members with a registered beta testing account - please have a look at the goodies available. AND TEST THEM :P
ID: 1207200 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1207297 - Posted: 18 Mar 2012, 1:23:36 UTC - in response to Message 1207200.  

Me being able to get it running at all is a great test!! Got it running and am now waiting for some work.


PROUD MEMBER OF Team Starfire World BOINC
ID: 1207297 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1207389 - Posted: 18 Mar 2012, 10:18:05 UTC - in response to Message 1206067.  

I wonder, do systems such as http://setiathome.berkeley.edu/show_host_detail.php?hostid=6311481 get similar messages telling them to update to a sensible version of BOINC? If not then they should!

They get an entry in the message log saying 'A newer version of BOINC is available', periodically.

The "A newer BOINC version is available" message was included in 5.2.3 onwards; 4.45 doesn't have it yet. And as can be easily deducted from that, what it doesn't have, it can't chirp, so it can't be added for him either. ;-)

Now, talking about wrong hosts... this one has so many wrongs. ;-)
ID: 1207389 · 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 1207397 - Posted: 18 Mar 2012, 10:36:03 UTC - in response to Message 1207389.  

I wonder, do systems such as http://setiathome.berkeley.edu/show_host_detail.php?hostid=6311481 get similar messages telling them to update to a sensible version of BOINC? If not then they should!

They get an entry in the message log saying 'A newer version of BOINC is available', periodically.

The "A newer BOINC version is available" message was included in 5.2.3 onwards; 4.45 doesn't have it yet. And as can be easily deducted from that, what it doesn't have, it can't chirp, so it can't be added for him either. ;-)

Ooops, sorry about that. BOINC v4.45 is before my time, grand-dad ;P - I think v5 was already out when I switched from SETI Classic, when the servers were turned off.

Now, talking about wrong hosts... this one has so many wrongs. ;-)

Like that blasted 296.10 NVidia driver, for a start.
ID: 1207397 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 34744
Credit: 261,360,520
RAC: 489
Australia
Message 1207400 - Posted: 18 Mar 2012, 10:41:48 UTC - in response to Message 1207389.  
Last modified: 18 Mar 2012, 10:42:50 UTC

Now, talking about wrong hosts... this one has so many wrongs. ;-)

I wonder if the problem with that 1 is the driver version (296.10) and the monitor being asleep.

Cheers.
ID: 1207400 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 1207403 - Posted: 18 Mar 2012, 10:48:45 UTC - in response to Message 1207397.  

I think v5 was already out when I switched from SETI Classic, when the servers were turned off.

I see here in Seti news, that it was 5.2.12 that was released to the public, around about 26 November 15, 2005. Seti Classic had its demise on December 15th, 2005, so it must've been 5.2.12 that you ran with at the time.

Too bad that the BOINC server had its couple of disk crashes last year, those gobbled up all those old versions and their release dates which were at that time still available from the download site and its mirrors. No more. :-(
ID: 1207403 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1207404 - Posted: 18 Mar 2012, 10:53:24 UTC - in response to Message 1207400.  
Last modified: 18 Mar 2012, 11:00:05 UTC

Now, talking about wrong hosts... this one has so many wrongs. ;-)

I wonder if the problem with that 1 is the driver version (296.10) and the monitor being asleep.

Cheers.

Don't think it's that Bug, when the Monitor goes to sleep, the cuda device disappears, then Stock Cuda app jumps into CPU fallback mode and the tasks end up with -177 maximum time exceeded errors,

the invalids here are SETI@Home Informational message -9 result_overflow's, most probably the GPU needs restarting or PSU isn't up to it,

Claggy
ID: 1207404 · 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 1207416 - Posted: 18 Mar 2012, 11:31:42 UTC - in response to Message 1207403.  

I think v5 was already out when I switched from SETI Classic, when the servers were turned off.

I see here in Seti news, that it was 5.2.12 that was released to the public, around about 26 November 15, 2005. Seti Classic had its demise on December 15th, 2005, so it must've been 5.2.12 that you ran with at the time.

Too bad that the BOINC server had its couple of disk crashes last year, those gobbled up all those old versions and their release dates which were at that time still available from the download site and its mirrors. No more. :-(

Just had a look on my server - which is still running, even though I had to replace the motherboard last month. I have v5.2.10 (downloaded 26 Nov 2005), and v5.2.13 (01 Dec 2005) - although by February 2006 I'd thrown them both out in favour of boinc-5.3.12.tx36

And what's this I see, on my daily driver? boinc_4.19_windows_intelx86.exe, downloaded 23 Feb 2005?

All of which reminds me - to get back on topic - that I have to downgrade that server from v6.10.58 to v5.something so I can test the service mode restart code for v5 in the new installer. Does your 'remove all traces of BOINC from my computer' script contain something to strip off the sandbox security group memberships from the data folder, so I can get back to a clean v5 service without losing work in progress, either here or for Einstein?
ID: 1207416 · Report as offensive
EdwardPF
Volunteer tester

Send message
Joined: 26 Jul 99
Posts: 389
Credit: 236,772,605
RAC: 374
United States
Message 1207445 - Posted: 18 Mar 2012, 13:20:55 UTC - in response to Message 1207403.  
Last modified: 18 Mar 2012, 13:27:26 UTC

Too bad that the BOINC server had its couple of disk crashes last year, those gobbled up all those old versions and their release dates which were at that time still available from the download site and its mirrors. No more. :-(



here are the dates I have for BOINC in my download archives ...



07/14/2004 12:05 PM 1,656,576 boinc_3.20_windows_intelx86.exe
08/27/2004 01:46 PM 1,668,080 boinc_4[1].05_windows_intelx86.exe
10/04/2004 09:21 PM 1,694,509 boinc_4.09_windows_intelx86.exe
10/12/2004 09:55 PM 1,697,170 boinc_4.12_windows_intelx86.exe
10/15/2004 06:03 PM 1,697,146 boinc_4.13_windows_intelx86.exe
02/13/2005 08:49 PM 1,697,670 boinc_4.19_windows_intelx86.exe
03/24/2005 01:08 PM 9,567,325 boinc_4.25_windows_intelx86.exe
08/20/2005 11:26 AM 392,791 boinc-445-sse.zip
08/20/2005 11:26 AM 365,258 boinc-445-mmx.zip
08/20/2005 11:27 AM 376,037 boinc-445-sse2.zip
10/27/2005 10:04 PM 10,514,521 boinc_5.2.5_windows_intelx86.exe
11/01/2005 03:28 PM 10,515,545 boinc_5.2.6_windows_intelx86.exe
11/17/2005 04:20 PM 10,524,249 boinc_5.2.7_windows_intelx86.exe
11/25/2005 11:29 PM 10,534,489 boinc_5.2.8_windows_intelx86.exe
11/27/2005 01:08 AM 10,542,170 boinc_5.2.12_windows_intelx86.exe
12/04/2005 10:16 AM 10,548,314 boinc_5.2.14_windows_intelx86.exe
02/17/2006 10:00 PM 10,546,778 boinc_5.2.13_windows_intelx86.exe
05/11/2006 11:29 PM 9,274,112 boinc_5.4.9_windows_intelx86.exe
01/03/2007 12:01 PM 9,304,320 boinc_5.4.11_windows_intelx86.exe
02/28/2007 01:30 PM 8,362,728 boinc_5.8.15_windows_intelx86.exe
09/06/2007 08:02 PM 6,967,200 boinc_5.10.20_windows_intelx86.exe
11/05/2007 02:41 PM 7,083,424 boinc_5.10.28_windows_intelx86.exe
12/09/2008 12:56 PM 7,127,456 boinc_6.2.19_windows_intelx86.exe
12/17/2008 10:11 PM 6,973,848 boinc_6.4.5_windows_intelx86.exe
12/19/2008 05:50 PM 3,579,540 Boinc_6.1.0.32_v5_release_x86.zip
09/15/2009 11:11 PM 6,694,304 boinc_6.6.36_windows_intelx86.exe
10/05/2009 10:48 AM 6,694,816 boinc_6.6.38_windows_intelx86.exe
12/06/2009 11:05 AM 7,354,272 boinc_6.10.18_windows_intelx86.exe
03/05/2010 08:43 PM 7,172,000 boinc_6.10.36_windows_intelx86.exe
04/07/2010 01:10 PM 7,177,120 boinc_6.10.43_windows_intelx86.exe
04/17/2010 10:28 PM 243,856 boinc and TM with AP Beta ATI13ATI.jpg
05/25/2010 01:51 AM 7,192,992 boinc_6.10.56_windows_intelx86.exe
08/04/2010 10:51 PM 7,216,544 boinc_6.10.58_windows_intelx86.exe
12/31/2011 05:07 PM 7,331,384 boinc_6.12.34_windows_intelx86.exe

I hope this helps

Ed F
ID: 1207445 · Report as offensive
Previous · 1 . . . 3 · 4 · 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.