Looking into Rosetta's source code?

Message boards : Number crunching : Looking into Rosetta's source code?

To post messages, you must log in.

AuthorMessage
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 71604 - Posted: 14 Nov 2011, 16:50:05 UTC

I am surprised how often the application "Rosetta Mini 3.17" crashed on my computer recently.
Is it possible to inspect corresponding source code for the affected executable files?
ID: 71604 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Rocco Moretti

Send message
Joined: 18 May 10
Posts: 66
Credit: 585,745
RAC: 0
Message 71605 - Posted: 14 Nov 2011, 18:24:17 UTC - in response to Message 71604.  

I am surprised how often the application "Rosetta Mini 3.17" crashed on my computer recently.
Is it possible to inspect corresponding source code for the affected executable files?


The source code to Rosetta (including the minirosetta boinc application) is available for free to academic and non-profit users through http://www.rosettacommons.org/software/. That might not help you exactly, as the version which is on Rosetta@home isn't the same as the released version.

Probably an easier course of action is to post whatever information about the crash you have here to the forums. Hopefully one of the other posters will recognize it as an existing issue and offer you a workaround, or perhaps one of the developers will be able to diagnose and fix it. (I can't make any promises on that regard, though.)
ID: 71605 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 71610 - Posted: 15 Nov 2011, 17:22:53 UTC - in response to Message 71605.  

That might not help you exactly, as the version which is on Rosetta@home isn't the same as the released version.

I would appreciate if I can look into the source files exactly for the currently used revision by a (distributed) content management system. How are the fulfilment chances for such a wish?
ID: 71610 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Rocco Moretti

Send message
Joined: 18 May 10
Posts: 66
Credit: 585,745
RAC: 0
Message 71611 - Posted: 15 Nov 2011, 19:36:02 UTC - in response to Message 71610.  

That might not help you exactly, as the version which is on Rosetta@home isn't the same as the released version.

I would appreciate if I can look into the source files exactly for the currently used revision by a (distributed) content management system. How are the fulfilment chances for such a wish?


Pretty slim, I'm sorry to say. The code that's being run on Rosetta@home corresponds to a fairly recent snapshot of the main Rosetta development branch. The official way of getting access to it would be to join a lab that's a member of RosettaCommons, and get your PI to vouch for your access to the code repository.

I believe that there's provisions for interested third parties to get access to the development source code as something like a "consultant" (as I understand it there's at least one member of the Foldit community working in that capacity). If you're serious and committed about volunteering to do testing/debugging of the Rosetta@home code, it might be worth emailing David Baker to volunteer in such a capacity, but I can't make any guarantees that anything will come of it. Keep in mind that David receives a large number of emails asking for things (so you might not get a response) and there's some effort and administrative/legal hurtles in setting up such an arrangement (so if it's just a temporary one-off interest, it's probably not going to be worth the hassle) and even then I don't know how disposed David would be to the arrangement for Rosetta@home (Foldit is a collaboration with Zoran Popović, and I believe it's mostly Zoran's lab that's working with the consultant.)
ID: 71611 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 71613 - Posted: 15 Nov 2011, 20:45:26 UTC - in response to Message 71605.  

Probably an easier course of action is to post whatever information about the crash you have here to the forums.

I assume that it will be very hard to find the real error reasons when so many computations fail. I imagine that automatic categorisation of computation failures would be needed to concentrate on interesting open issues, wouldn't it?

Hopefully one of the other posters will recognize it as an existing issue and offer you a workaround, or perhaps one of the developers will be able to diagnose and fix it.

I would be glad if somebody will recognize useful patterns.

Issue candidates:
1. Task 463437031: Compute error
Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Out Of Memory (C++ Exception)

2. Task 462767468: Compute error
SIGBUS: bus error

3. Task 462746389: Validate error

4. Task 463638378: Compute error
Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation
ID: 71613 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 71614 - Posted: 15 Nov 2011, 21:28:53 UTC - in response to Message 71611.  

The code that's being run on Rosetta@home corresponds to a fairly recent snapshot of the main Rosetta development branch.

Thanks for your clarification.

I believe that there's provisions for interested third parties to get access to the development source code as something like a "consultant" (...).

I appreciate your explanation for the current situation.

Are there any chances to improve the accessibility to the really used program sources for interested free software developers like me?

Do the involved universities care for "Open Access" (with implementation documentation and source files)?
ID: 71614 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
jmowens

Send message
Joined: 28 Nov 11
Posts: 1
Credit: 0
RAC: 0
Message 71665 - Posted: 28 Nov 2011, 19:02:16 UTC - in response to Message 71605.  

I am surprised how often the application "Rosetta Mini 3.17" crashed on my computer recently.
Is it possible to inspect corresponding source code for the affected executable files?


The source code to Rosetta (including the minirosetta boinc application) is available for free to academic and non-profit users through http://www.rosettacommons.org/software/. That might not help you exactly, as the version which is on Rosetta@home isn't the same as the released version.

Probably an easier course of action is to post whatever information about the crash you have here to the forums. Hopefully one of the other posters will recognize it as an existing issue and offer you a workaround, or perhaps one of the developers will be able to diagnose and fix it. (I can't make any promises on that regard, though.)


For clarification: The license links available at http://www.rosettacommons.org/software/ are academic and commercial licenses obtained through UW Center for Commercialization for use of the Rosetta software, and going through this process facilitates downloading the software for the individual licensee's use.

Access to the Rosetta source code repository is obtain through separate channel(s) not involving these licenses.

Best regards,
Jennifer Owens
UW Center for Commercialization
jmowens(at)uw(dot)edu
ID: 71665 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 71745 - Posted: 4 Dec 2011, 20:40:16 UTC - in response to Message 71611.  

The code that's being run on Rosetta@home corresponds to a fairly recent snapshot of the main Rosetta development branch.

How many software developers have got direct access to the original source files for Rosetta's applications?

Would you like to consider further ways to support easier cooperation and improvements for a higher contribution probability?
ID: 71745 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 71835 - Posted: 20 Dec 2011, 14:00:33 UTC - in response to Message 71611.  

If you're serious and committed about volunteering to do testing/debugging of the Rosetta@home code, it might be worth emailing David Baker to volunteer in such a capacity, but I can't make any guarantees that anything will come of it.

Have you got a more direct contact to him?

Are you interested in further clarification of access ways to the original source files that drive software like the application "Rosetta Mini 3.19" currently?
ID: 71835 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Keith E. Laidig
Volunteer moderator
Project developer
Avatar

Send message
Joined: 1 Jul 05
Posts: 154
Credit: 117,189,961
RAC: 0
Message 72008 - Posted: 8 Jan 2012, 14:50:27 UTC

Markus,

Sorry, but no, you can't get the code that is being run in R@H at the moment. It is
in active development and the scientists do not release all development branches of the code
tree to the public prior to publication of their work.

You could, in principle, get a copy of the base code through the UW - as you know.

Thanks for your interest.

ID: 72008 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 72029 - Posted: 9 Jan 2012, 11:11:00 UTC - in response to Message 72008.  

You could, in principle, get a copy of the base code through the UW - as you know.

Jennifer Owens informed me that she can not help me with the requested source code access so far. Who can decide from your team to provide/grant a look into the used source files?

Would you like to turn your offer into free software?
ID: 72029 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Aegis Maelstrom

Send message
Joined: 29 Oct 08
Posts: 61
Credit: 2,137,555
RAC: 0
Message 72110 - Posted: 13 Jan 2012, 17:40:22 UTC - in response to Message 72029.  

Regarding the general question of ability to take a look into the Rosetta source code - have you thought of external "consulting" from the volunteering programmers from the BOINC or Free/Libre/Open Source Software community?

I do realize there can be many secrets and fears of being copied involved - but additional IT resources could be very potent: both in generating new and higher quality, reviewed code as well as higher reliability and recognition within our communities.

For instance, as an active member of BOINC@Poland team's community and forum, I see questions regarding Rosetta's efficiency, possibility and gains from 64-bit version, use of additional instructions and registers like SSEs and AVX, general recompiling etc. etc. quite often. These are understandable concerns which we have sometimes a hard time to answer.

My teammates do run a few (however smaller) BOINC Projects and introduced a few improvements to some other existing ones so there is some expertise even within a single team, not mentioning the broad BOINC/FLOSS communities.

I remember seeing on this forum a very brief remark of a volunteer who saw some actual code and I thought it was a positive experience for both sides. Even should you need to show only some less biology specific part of the code, use some NDAs etc. etc., maybe you could coin some mechanism to share the knowledge, gain some good PR and grasp improvements?

Best regards,
a.m.
ID: 72110 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile robertmiles

Send message
Joined: 16 Jun 08
Posts: 1217
Credit: 13,359,660
RAC: 780
Message 72132 - Posted: 14 Jan 2012, 21:41:36 UTC
Last modified: 14 Jan 2012, 21:42:25 UTC

I've read elsewhere that they charge for a commercial version of the source code. That appears likely to provide a major part of their income, so do you expect them to make it free?
ID: 72132 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 72141 - Posted: 15 Jan 2012, 12:50:20 UTC - in response to Message 72132.  

I've read elsewhere that they charge for a commercial version of the source code.

Yes. - This university is also interested in the commercialisation of "its assets".

That appears likely to provide a major part of their income, so do you expect them to make it free?

Support services and education can also be sold for free software, can't it?

Would you like to benefit from more freedom in this science and software domain?
ID: 72141 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Markus Elfring

Send message
Joined: 10 Jun 06
Posts: 17
Credit: 3,610,273
RAC: 0
Message 72142 - Posted: 15 Jan 2012, 13:50:28 UTC - in response to Message 72110.  

... - have you thought of external "consulting" from the volunteering programmers from the BOINC or Free/Libre/Open Source Software community?

I am volunteering for some other applications already.

I do realize there can be many secrets and fears of being copied involved

I hope that such fears can be reduced.

- but additional IT resources could be very potent: both in generating new and higher quality, reviewed code as well as higher reliability and recognition within our communities.

I would also appreciate if Rosetta's applications can benefit from corresponding effects.

These are understandable concerns which we have sometimes a hard time to answer.

Do you get any answers for the involved issues from the project leaders?

I remember seeing on this forum a very brief remark of a volunteer who saw some actual code and I thought it was a positive experience for both sides.

Would you like to improve the experience a bit more in this direction?
ID: 72142 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : Looking into Rosetta's source code?



©2022 University of Washington
https://www.bakerlab.org