Advanced search

Message boards : News : [dna] new DNA@Home runs up and going

1 · 2 · Next
Author Message
Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6048 - Posted: 18 Jan 2016, 5:30:10 UTC

It looks like everything is coming back just fine with the new application versions, so I've started up a batch of fresh runs which will mean a lot of new workunits. Happy crunching!

[VENETO] boboviz
Send message
Joined: 10 Aug 14
Posts: 91
Combined Credit: 387,515
DNA@Home: 184,990
SubsetSum@Home: 1,636
Wildlife@Home: 200,890
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

    
Message 6051 - Posted: 19 Jan 2016, 20:06:03 UTC - in response to Message 6048.

Welcome back!!!

P.s.
What's the difference with old app??

mugnaio [TNAA]
Send message
Joined: 11 Aug 14
Posts: 8
Combined Credit: 721,094
DNA@Home: 105,856
SubsetSum@Home: 28,660
Wildlife@Home: 586,578
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

      
Message 6052 - Posted: 19 Jan 2016, 20:21:33 UTC - in response to Message 6051.

What's the difference with old app??


same question here... The new one is built linking glibc 2.14 and that prevents WU from running on debian wheezy, while release 0.50 was running fine.

TIA
Andrea

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6053 - Posted: 19 Jan 2016, 21:13:07 UTC - in response to Message 6052.
Last modified: 19 Jan 2016, 21:13:37 UTC

What's the difference with old app??


same question here... The new one is built linking glibc 2.14 and that prevents WU from running on debian wheezy, while release 0.50 was running fine.

TIA
Andrea


That's really strange. I used the exact same VMs that I compiled the previous apps with, so there shouldn't be any difference there. Is this the 32-bit linux app?

The new app is printing out additional information we need to perform some additional analysis of the results we're getting. We'll also be able to use it to update the DNA@Home progress pages with some additional visualizations so we can better track how well the searches are progressing.

We're also whipping up some new data sets for runs that will be done on the mouse genome as opposed to the human genome (which the current data sets are from).

GlennG
Send message
Joined: 19 Aug 14
Posts: 6
Combined Credit: 32,628,601
DNA@Home: 8,447,080
SubsetSum@Home: 23,435,613
Wildlife@Home: 745,908
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

      
Message 6066 - Posted: 20 Jan 2016, 13:15:21 UTC - in response to Message 6053.

Travis,

Gibbs_0.51_x86_64-pc-linux-gnu

0.51 is requiring newer GCC and G++ libraries than 0.50 required, and as such many older machines are now incompatible. This is same problem that occurred with the last update to subset last year.
____________

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6073 - Posted: 20 Jan 2016, 16:18:25 UTC - in response to Message 6066.

Travis,

Gibbs_0.51_x86_64-pc-linux-gnu

0.51 is requiring newer GCC and G++ libraries than 0.50 required, and as such many older machines are now incompatible. This is same problem that occurred with the last update to subset last year.


Any chance you could let me know what would be a good version of the GCC and G++ libraries to link? I can try and get an older compiler installed to link against.

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6074 - Posted: 20 Jan 2016, 17:31:02 UTC - in response to Message 6066.
Last modified: 20 Jan 2016, 17:31:12 UTC

Travis,

Gibbs_0.51_x86_64-pc-linux-gnu

0.51 is requiring newer GCC and G++ libraries than 0.50 required, and as such many older machines are now incompatible. This is same problem that occurred with the last update to subset last year.


So I'm currently compiling with gcc 4.8 (which is already quite old). I can't even get any of the gcc 3.x versions to work on the virtual machine... I'm not sure if there's going to be a good solution here as those versions of gcc are ancient, and I'm not sure if I'll even be able to get BOINC to compile with them (or my linux VM to even run them).

mugnaio [TNAA]
Send message
Joined: 11 Aug 14
Posts: 8
Combined Credit: 721,094
DNA@Home: 105,856
SubsetSum@Home: 28,660
Wildlife@Home: 586,578
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

      
Message 6075 - Posted: 20 Jan 2016, 19:48:18 UTC - in response to Message 6073.

Any chance you could let me know what would be a good version of the GCC and G++ libraries to link? I can try and get an older compiler installed to link against.


Not so easy to give a good guess; Debian, which is quite conservative on package upgrades, provides these:
- current stable release (jessie, released in april 2015): gcc-4.9, glibc 2.19
- old stable release (wheezy, released in may 2013): gcc-4.7, glibc 2.13
- old old stable release (squeeze, released in february 2011): gcc-4.4, glibc 2.11
while these should be the packages in Ubuntu:
- wily werewolf, released in october 2015: gcc-5.2, glibc 2.21
- trusty tahr, released in april 2014: gcc-4.8, glibc 2.19
- precise pangolin, release in april 2012:gcc-4.6, glibc 2.15
I took a look also at an old fedora 12 released in 2010: gcc-4.4, glibc 2.11
and a Centos 6.5 released in july 2011: gcc-4.4, glibc 2.12
(release versions taken from distribution repositories, any mistake is mine)

That said, it's all up to you :)

my 2c is that rel. 0.50 was working fine, or at least nobody complained on this forum, so the same settings would be fine.

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6076 - Posted: 20 Jan 2016, 20:10:24 UTC - in response to Message 6075.

Any chance you could let me know what would be a good version of the GCC and G++ libraries to link? I can try and get an older compiler installed to link against.


Not so easy to give a good guess; Debian, which is quite conservative on package upgrades, provides these:
- current stable release (jessie, released in april 2015): gcc-4.9, glibc 2.19
- old stable release (wheezy, released in may 2013): gcc-4.7, glibc 2.13
- old old stable release (squeeze, released in february 2011): gcc-4.4, glibc 2.11
while these should be the packages in Ubuntu:
- wily werewolf, released in october 2015: gcc-5.2, glibc 2.21
- trusty tahr, released in april 2014: gcc-4.8, glibc 2.19
- precise pangolin, release in april 2012:gcc-4.6, glibc 2.15
I took a look also at an old fedora 12 released in 2010: gcc-4.4, glibc 2.11
and a Centos 6.5 released in july 2011: gcc-4.4, glibc 2.12
(release versions taken from distribution repositories, any mistake is mine)

That said, it's all up to you :)

my 2c is that rel. 0.50 was working fine, or at least nobody complained on this forum, so the same settings would be fine.


I'll try and get a debian VM up and going and see if I can compile things with an older version of GCC.

The current version is compiled with GCC 4.8, so I'm not quite sure how old I need to go. To be honest I don't remember what version 0.50 was compiled with because it was so long ago.

mugnaio [TNAA]
Send message
Joined: 11 Aug 14
Posts: 8
Combined Credit: 721,094
DNA@Home: 105,856
SubsetSum@Home: 28,660
Wildlife@Home: 586,578
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

      
Message 6077 - Posted: 20 Jan 2016, 20:43:38 UTC - in response to Message 6076.

I'll try and get a debian VM up and going and see if I can compile things with an older version of GCC.

The current version is compiled with GCC 4.8, so I'm not quite sure how old I need to go. To be honest I don't remember what version 0.50 was compiled with because it was so long ago.


google told me to use readelf -wi on that version, and this is what it says:
GNU C++ 4.4.6 20110731 (Red Hat 4.4.6-3)

Hubbie
Send message
Joined: 26 Jan 15
Posts: 6
Combined Credit: 126,669
DNA@Home: 100,488
SubsetSum@Home: 26,181
Wildlife@Home: 0
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

    
Message 6086 - Posted: 21 Jan 2016, 17:39:49 UTC - in response to Message 6048.

Hi Travis,

there seems to be a buffer limit in place of 4 workunits per core. Would it be possible to raise that number to (at least) 10 or so? Most of the workunits take just 30 seconds to about 5 minutes to process, which means I'm running out of work frequently at the moment. Thanks!

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6089 - Posted: 21 Jan 2016, 19:26:28 UTC - in response to Message 6086.

Hi Travis,

there seems to be a buffer limit in place of 4 workunits per core. Would it be possible to raise that number to (at least) 10 or so? Most of the workunits take just 30 seconds to about 5 minutes to process, which means I'm running out of work frequently at the moment. Thanks!


I think you've gotten unlucky! Some take up to 3 hours (at least on my mac pro, which is pretty beefy). Due to how the project works I'm reluctant to send out too many DNA workunits at the same time, as new workunits are dependent on previous ones being returned - so if many are buffered up that's going to make work availability as a whole suffer.

On a positive note I should have a load of subset sum workunits to fill in the gap out tonight.

Alessio Susi
Avatar
Send message
Joined: 10 Mar 15
Posts: 12
Combined Credit: 69,338
DNA@Home: 42,441
SubsetSum@Home: 10,466
Wildlife@Home: 16,431
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 0

      
Message 6108 - Posted: 25 Jan 2016, 20:19:14 UTC

I noticed the same thing. Some tasks take also 10-20 seconds to terminate and I receive 10-15 points. It would be nice to have longer tasks or more ones per core
____________
MoBo: MSI X370 Gaming Plus
CPU: AMD Ryzen 7 1700, 8 core / 16 thread 3.8 GHz
GPU: AMD Radeon Sapphire RX 480 4GB Nitro+
RAM: 2x8 GB Kingston HyperX Fury 2933 MHz
Windows 10 Pro

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6119 - Posted: 28 Jan 2016, 20:42:52 UTC - in response to Message 6108.

I noticed the same thing. Some tasks take also 10-20 seconds to terminate and I receive 10-15 points. It would be nice to have longer tasks or more ones per core


Unfortunately we're doing different sized runs to test different parameters to the algorithm. So there's most likely going to be a mix for awhile.

Travis Desell
Volunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 16 Jan 12
Posts: 1790
Combined Credit: 2,265,607
DNA@Home: 293,563
SubsetSum@Home: 349,212
Wildlife@Home: 1,622,832
Wildlife@Home Watched: 212,926s
Wildlife@Home Events: 51
Climate Tweets: 21
Images Observed: 710

              
Message 6120 - Posted: 28 Jan 2016, 20:42:52 UTC - in response to Message 6108.
Last modified: 28 Jan 2016, 20:43:21 UTC

I noticed the same thing. Some tasks take also 10-20 seconds to terminate and I receive 10-15 points. It would be nice to have longer tasks or more ones per core


Unfortunately we're doing different sized runs to test different parameters to the algorithm. So there's most likely going to be a mix for awhile. It would be nice if they could be all the same (and a longer size), but the science really doesn't allow it at the moment. Hopefully down the line we'll get there though!

Jesse Viviano
Send message
Joined: 29 Nov 14
Posts: 2
Combined Credit: 4,595,205
DNA@Home: 43,493
SubsetSum@Home: 61,402
Wildlife@Home: 4,490,310
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 117
Images Observed: 0

        
Message 6122 - Posted: 31 Jan 2016, 17:14:42 UTC

I think that the bug that causes results to be different if the program is shut down and then started back up is still present. See http://csgrid.org/csg/workunit.php?wuid=4898825 for one work unit where my result at http://csgrid.org/csg/result.php?resultid=10285824 was marked as invalid because I was forced to shut down BOINC and then restart it in the middle of processing this work unit.

kain
Send message
Joined: 24 Jan 15
Posts: 7
Combined Credit: 282,884
DNA@Home: 282,884
SubsetSum@Home: 0
Wildlife@Home: 0
Wildlife@Home Watched: 0s
Wildlife@Home Events: 0
Climate Tweets: 0
Images Observed: 1

  
Message 6374 - Posted: 28 Aug 2016, 18:52:54 UTC

Is this project dead? It certainly looks like this...

Profile wichtig77
Send message
Joined: 24 Feb 15
Posts: 4
Combined Credit: 338,972
DNA@Home: 18,392
SubsetSum@Home: 106,423
Wildlife@Home: 214,157
Wildlife@Home Watched: 528,531s
Wildlife@Home Events: 436
Climate Tweets: 520
Images Observed: 6,062

              
Message 6376 - Posted: 13 Sep 2016, 20:16:51 UTC

looks like the whole side is dead!
very sad
____________

Profile Steve Hawker*
Send message
Joined: 8 Apr 13
Posts: 134
Combined Credit: 829,896
DNA@Home: 11,932
SubsetSum@Home: 299,708
Wildlife@Home: 518,257
Wildlife@Home Watched: 5,541,577s
Wildlife@Home Events: 2,169
Climate Tweets: 8,579
Images Observed: 55

              
Message 6377 - Posted: 14 Sep 2016, 17:15:24 UTC - in response to Message 6376.
Last modified: 14 Sep 2016, 17:15:37 UTC

looks like the whole side is dead!
very sad

~850 people ran SSS in the last 24 hours. That hardly qualifies as dead.
http://csgrid.org/csg/server_status.php

Of course you can always classify some tweets or watch some videos. I expect that eventually there might be an app for both of these but without a significant data set to compare against, probably not worth it. So if you want more apps to crunch, help out with the tweets and the videos.

Travis has already said that DNA is waiting for some data sets. That was after school closed down for the summer and we're only September. Sometimes science doesn't run as fast as you'd like...
http://csgrid.org/csg/forum_thread.php?id=2102&postid=6308

Profile DoctorNow
Avatar
Send message
Joined: 31 Jan 12
Posts: 328
Combined Credit: 2,028,390
DNA@Home: 94,617
SubsetSum@Home: 162,940
Wildlife@Home: 1,770,832
Wildlife@Home Watched: 1,977,485s
Wildlife@Home Events: 589
Climate Tweets: 128
Images Observed: 2,614

              
Message 6378 - Posted: 15 Sep 2016, 11:30:29 UTC - in response to Message 6377.
Last modified: 15 Sep 2016, 11:39:39 UTC

~850 people ran SSS in the last 24 hours. That hardly qualifies as dead.
http://csgrid.org/csg/server_status.php

That's not the issue here. I also feel as if this project is dead because:
There is no update to anything here anymore - no development, no word, nothing.
I asked for an update to the image reviews about a half year ago - counts and badges were announced but nothing happened since then...
I have given up watching anything because of this. >:(
____________
Wildlife & SubsetSum Badges-creator ;-)
Mod/Guru @ the german team BOINC@Heidelberg
My BOINC-Stats / My Badges

1 · 2 · Next
Post to thread

Message boards : News : [dna] new DNA@Home runs up and going