Advanced search

Message boards : Wish list : Why are GPUGrid applications proprietary?

Author Message
ahj
Avatar
Send message
Joined: 25 Sep 10
Posts: 10
Credit: 31,143,381
RAC: 0
Level
Val
Scientific publications
watwatwatwatwatwatwatwat
Message 33609 - Posted: 25 Oct 2013 | 2:29:05 UTC
Last modified: 25 Oct 2013 | 2:32:28 UTC

Greetings

First of all I'd like to thank the GPUGrid team for pushing the boundaries of molecular dynamics to great new heights these past few years. Although I haven't crunched for GPUGrid (or any BOINC project, for that matter) in a long time, it was my favorite distributed computing project - but now, alas, I was disappointed to learn that the molecular dynamics applications used by GPUGrid are proprietary. Because of this, I can no longer, with good conscience, recommend or return to the GPUGrid project.

It is of my opinion that all software that serves a function ought to be free - not free as in zero price, but free as in freedom - i.e. the source code should be liberated so that it can be peer reviewed and peer assessed. The scientific method itself is based on the concepts of transparency and reproducibility - and I think for a worthy project like GPUGrid, to make its scientific applications a secret, is a real shame.

So I propose a question to the GPUGrid team: why don't you follow the trailblazers of free, liberated scientific software such as the Einstein, SETI and Milkyway @home projects? All of these projects license their science applications under a free software license (e.g the GNU GPL or MIT license).

Scientific progress is hurt when certain tools used to postulate and test new ideas are kept secret. I hope the GPUGrid team will seriously reconsider re-licensing their proprietary applications under a liberal license.

Regards

An old GPUGrid cruncher.

Stefan
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 5 Mar 13
Posts: 348
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 33611 - Posted: 25 Oct 2013 | 8:55:21 UTC - in response to Message 33609.

Hey ahj,

the simulations we do are not proprietary and there are plans to share the data with the scientific community aside from just the publications.
Now if you are talking about the simulation software, this is not what GPUGrid scientists are developing. We just use ACEMD because it was written by Gianni and Matt and in that sense it is convenient.
ACEMD is not open source as Acellera is selling it, but seriously, you could do the exact same work with any other open source MD software.

So considering that we are not developing simulation software but rather use it to investigate biological systems, and considering that you could very easily just swap ACEMD out for another software with no significant difference, I would say that it doesn't matter so much that the tool we use to do them is closed source.
On the other hand, simulation methodologies and biological systems are published by the group on a regular basis.

My point is, AFAIK if you want to do what we do for free and open source you can do it.

ahj
Avatar
Send message
Joined: 25 Sep 10
Posts: 10
Credit: 31,143,381
RAC: 0
Level
Val
Scientific publications
watwatwatwatwatwatwatwat
Message 33613 - Posted: 25 Oct 2013 | 11:18:26 UTC - in response to Message 33611.

Hi Stefan, thanks for your quick reply.

I understand exactly that you release and publish your results; and I can understand why you would want to use an in-house MD program rather than a third party one.

The only issue I have is regarding the computational application itself - ACEMD. You mentioned that the only reason it isn't open source is because it is being sold by Acellera; but perhaps there is a misunderstanding. There is no need to make a program proprietary if you wish to sell it; free software licenses such as the GPL/MIT licenses protect and guarantee the right for an author to sell his/her program.

And that's what the guys at Milkyway and Einstein have done - the applications themselves are released as free software.

Just out of curiosity, would it be possible to use free software instead, say for instance, Gromacs and/or OpenMD?

I look at the work done by you and the GPUGrid team with the utmost admiration and respect, but if the tool - ACEMD - that is used to perform these important computations is kept secret, it hurts the scientific/high performance computing/molecular dynamics community who could study, share, modify and improve the code.

I know it's unlikely that ACEMD will ever be liberated due to the arrangement with Acellera, but I think that if the GPUGrid team liberated it, it would make a huge contribution to the MD community; perhaps as important as the other free software programs Gromacs, AMBER and OpenMD.

Regards

Jozef J
Send message
Joined: 7 Jun 12
Posts: 112
Credit: 1,118,845,172
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwat
Message 33627 - Posted: 26 Oct 2013 | 18:07:10 UTC - in response to Message 33613.

NO..!!GPUGRID is a private project all I calculate is investing in a private Spanish company. ... you are very naive .. I do not get how this project came under the wing boinc ... therefore we will leave nobody, but nobody wants to answer you ..

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 33631 - Posted: 26 Oct 2013 | 22:48:04 UTC - in response to Message 33627.

NO..!!GPUGRID is a private project all I calculate is investing in a private Spanish company. ... you are very naive .. I do not get how this project came under the wing boinc ... therefore we will leave nobody, but nobody wants to answer you ..

It's actually the other way around: the 'Spanish Company' (Acellera) investing in GPUGrid, by letting us use their proprietary molecular dynamic simulation software for producing results which are in the public domain. In exchange we help them to develop their MD software by using it. It's a win-win scenario :).

Stefan
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 5 Mar 13
Posts: 348
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 33648 - Posted: 27 Oct 2013 | 22:30:33 UTC - in response to Message 33631.
Last modified: 27 Oct 2013 | 22:50:09 UTC

Not to mention the work Matt (MJH) does for this project fixing all the problems even though it's not his job. Acellera is (afaik) getting quite little out of GPUGrid (spreading the name and maybe a few bugs?), so the cooperation is quite good for us considering the help we get.

@ahj: I am sure they have reviewed the option to make the code open but in the end it is Acellera's decision, not the group's.
But to tell you the truth, I don't know how much of a gain there would be for the MD community in making it open source. Based on the way other MD software advances ACEMD is probably on the same level as the other software. And we are not per-se against using open software. We are actually looking into using other open software to expand GPUGrid even more.
I understand your concern but at the moment it is more for our advantage to do it this way I guess. But it might change in the future.
In my totally personal opinion I don't see any harm being done.

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2705
Credit: 1,311,122,549
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 33673 - Posted: 30 Oct 2013 | 12:26:59 UTC

Ahj, you're right that it's the best scientific practise to publish source code. But, as soon as commercial interests are concerned it's a bit naive to trust the protection of a liscence alone: "They not allowed to use our software for their commercial purposes, they have to buy it for that. I'm sure they'd never just take it silently."

MrS
____________
Scanning for our furry friends since Jan 2002

Post to thread

Message boards : Wish list : Why are GPUGrid applications proprietary?

//