Message boards : Rosetta@home Science : To David Baker, David Kim, Jack, & Other Admin
Previous · 1 · 2
Author | Message |
---|---|
David Baker Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 17 Sep 05 Posts: 705 Credit: 559,847 RAC: 0 |
I just wanted to say thank you for the great communication that you seem to be giving your participants here. It seems that you stay in the forums and chat well, and keep people up to date and let people know what is going on! Thank you Aaron and others for this positive feedback--it is really great! This has been a lot of fun for me and the others on the team. Part of our job as scientists is to communicate what we are trying to learn to other people. for example, right now I'm in new york city (hence this response will be short!) and will soon be leaving for seminars at princeton (tomorrow), yale (tuesday), and a group of universities in nyc before returning to seattle on wednesday. I don't like travelling (I hate planes and airports!) and it is great to be able to communicate with all of you in this way instead--maybe I can skip the planes entirely soon.. Another reason we are trying to make this as fun for everybody as possible is that the single major stumbling block to progress for us is computing power--and you are making it possible for us to approach problems we never could have solved before. The problem with most biology related computation has been that the results were just not very accurate, and so computing power didn't really matter so much. The big news in our two recent Science papers (one on structure prediction, one on modelling biological macromolecules more generally), is that for the first time it is really possible to compute things like protein structures and itneractions accurately, but that it takes lots of computing power. Hence, you are all making possible what we hope will be a revolution in biology--the transformation from a purely experimental science to one in which computation plays a major role in solving problems. In other words, you are getting more feedback from us probably because we need you more! |
Fuzzy Hollynoodles Send message Joined: 7 Oct 05 Posts: 234 Credit: 15,020 RAC: 0 |
And that's exactly why participating in distributed computer projects is so fun! I love being able to let my computer be a part of research, whether it's particle physics, or biology, in tasks, one would have considered impossible only a few years ago! That makes it worth it! [b]"I'm trying to maintain a shred of dignity in this world." - Me[/b] |
Plum Ugly Send message Joined: 3 Nov 05 Posts: 24 Credit: 2,005,763 RAC: 0 |
Well I got spoiled with the fantastic job Keith did with comunicating how and what the project was doing.I hope David keeps doing us the same way here. Thanks Kel for the explaination too. It gives us fadders more info into Rosetta. But I'm still a little concerned about the servers. |
David E K Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1480 Credit: 4,334,829 RAC: 0 |
We assure you all that there is nothing to worry about with our servers. We will scale as necessary and maintenance down times should not extend past the amount of time your clients can process work without communication with the server. Our lab consists of over 500 production nodes (see images on Keith's web page) but accounts for a fraction of the throughput that this project provides, so we are definitely focusing resources towards this project. |
Keith E. Laidig Volunteer moderator Project developer Send message Joined: 1 Jul 05 Posts: 154 Credit: 117,189,961 RAC: 0 |
But I'm still a little concerned about the servers. Well, don't worry too much about the hardware. We've been planning on 'beefin' up' our infrastructure to address growth in the project and we've tasked a substantial budget towards keeping things running smoothly. And, as I'm the laziest man in the world, I'm gonna build this system up to be bullet-proof and I'm not worried how much of DB's $$$ I've got to spend to make it happen; load balancing over multiple servers, multiple, mirrored database servers, huge memory backend machines, we're even looking at how we might set up mirrored installations at our two different computer centers to ensure that there is always a timely response from us here at R@H. DB likely doesn't share my glib disregard for his budgets, so I imagine they'll be some comprimise on things - but we'll still build something mighty impressive to match up appropriately to your contributions. You see, you (and/or your computers) are much more important to us than you might appreciate. We're betting the house on you. To be honest, I had thought that we'd get most of our members from the local UW campus. I had thought we'd slowly build up the project over the academic year and, thus, have time to design infrastructure to address the slowly growing user-base. 'You silly naive, boy', you're likely saying now.... I was unaware there were so many out there ready to contribute. You gals and guys have kinda been freaking me out. I mean, who would have suspected that there are so many computer/science-geeks out there! I think it is good that so many wish to be involved, but based upon the average response to my discussions at the local pub I would have extrapolated interest in computational biophysics world wide to include around fourteen people in total. Clearly, I was mistaken. This probably belongs in the 'Cruching' board, but I'll start it here for consistency sake...here's what we are seeing: Specifically, the primary bottle-neck we've seen to date is the modest, yet persistent I/O load we see between the database engine and the InnoDB files. This hasn't proven to be much of an operational problem to date, but we've started to address the issues (which were the outages you might have noticed last week). We doubled the memory in the database server in the hopes of caching more of the datastructure and thereby reducing the amoung that the database 'page faults' to disk. DK also started tuning the behavior of the database engine and massaging the datastructure a bit. We've been in contact with SETI@HOME and Einstein@home to see how they handle things on the hardware end and we're looking at incorporating their experience in our hardware designs. Lastly, if there is interest - and I won't take it personally if there isn't - I can outline the path and methods we undertake to build up the R@H servers. For example, I'll be doing some testing on - hopefully - ultra-fast fileservers in the next little bit for the next generation fileserver and can let you know what I'm finding.... |
FZB Send message Joined: 17 Sep 05 Posts: 84 Credit: 4,948,999 RAC: 0 |
that looks nice, lot's of 2U servers... actually looks like the place i want to call home :D @kel: depends on your time but making your experience with configuring the boinc server infastructure available might help future boinc projects which sounds like a good idea to me (though i bet most projects will not build a top notch fileserver due to budget ;) ) -- Florian www.domplatz1.de |
AnRM Send message Joined: 18 Sep 05 Posts: 123 Credit: 1,355,486 RAC: 0 |
Quote.....Lastly, if there is interest - and I won't take it personally if there isn't - I can outline the path and methods we undertake to build up the R@H servers. For example, I'll be doing some testing on - hopefully - ultra-fast fileservers in the next little bit for the next generation fileserver and can let you know what I'm finding....[/quote] >I think that would be great! SETI has a 'Technical News' header which gives more detail on the techie side of things. Something like that might be helpful. Thanks, Keith, for taking the time to enlighten us on your plans. The people at E@H have developed the most reliable BOINC project (of the five we run anyway) and I'm reassured when you mention that you will be talking to them. You guys are doing a fantastic job for a new BOINC project (very few glitches, really) and it bodes well for R@H. IMHO, communication, stability and intresting science are the ingredients for a successful projects....and you have most of these in spades....Cheers, Rog. |
Paul D. Buck Send message Joined: 17 Sep 05 Posts: 815 Credit: 1,812,737 RAC: 0 |
Kel, This is also the kind of information that SHOULD be in the documentation ... Which is why I have been trying to also get projects interested in a cross-project repository of BOINC information. Which at the moment seems to be the Wiki ... the biggest advantage to it is that it can be edited by all allowed editors so that the content is not dpendent on one or two people (at least in theory). I HAD hoped to be working on the development side of the material,but, well, I am not working on much at all these days ... Anyway, one of the things I had hoped to add to the development side was more notes on how to do a simple installation (one or two servers) then how to "grow" that in a logical way ... so, either way I am veery interested in the information as I think this is extrodinarily important and is not documented at all ... |
dgnuff Send message Joined: 1 Nov 05 Posts: 350 Credit: 24,773,605 RAC: 0 |
Never mind - addressed far better by others. BTW, is there a "delete post" option anywhere? |
Essee_Ar Send message Joined: 7 Nov 05 Posts: 10 Credit: 351,015 RAC: 0 |
We assure you all that there is nothing to worry about with our servers. We will scale as necessary and maintenance down times should not extend past the amount of time your clients can process work without communication with the server. Our lab consists of over 500 production nodes (see images on Keith's web page) but accounts for a fraction of the throughput that this project provides, so we are definitely focusing resources towards this project. Nice equipment! Sooooo.... what do you do with your castoffs? :) Does equipment that has been replaced turn into "crunchers"? Ya gotta know there is more than a few hardware junkies in here! :) Scott |
Keith E. Laidig Volunteer moderator Project developer Send message Joined: 1 Jul 05 Posts: 154 Credit: 117,189,961 RAC: 0 |
Sooooo.... what do you do with your castoffs? :) Sorry Scott, we run all of our gear right into the ground. I'm the cheapest bastard around and I won't shut something off until it well and truely dead. After that I canabalize everything, leaving only a empty 'husk' which the UW then sells in huge batches to...I don't know. -KEL |
Fuzzy Hollynoodles Send message Joined: 7 Oct 05 Posts: 234 Credit: 15,020 RAC: 0 |
You're wellcome, Kel! :-) And I think I can speak for everybody here, when I say that we're delighted! Yes, as said on our team site, "We're a friendly bunch who have gathered together from all corners of the globe to take part in a fascinating project, to try and find the answer to the ultimate question . ." So thank you for letting us! :-) [b]"I'm trying to maintain a shred of dignity in this world." - Me[/b] |
nasher Send message Joined: 5 Nov 05 Posts: 98 Credit: 710,554 RAC: 2,975 |
to be honest when you start geting the larger influx of people from FaD you will find that we are for the most part intrested in reading anything if you look at the Find a drug boards and realize the posts there have droped to about 1/100th what they used you will see we like information. Yes so far in my serch for the DC project i want to run Rosetta has been the Best i could come up for me. (corse i will have a few others setup on boinc) I am glad to see the responses from the scientists here and hope we get more (ok sorry my spelling sucks) thought of adding a spellchecker to the messageboards? oh well it looks like a good project to me please give us whatever information you want to.. heck on Fad there was even a few blogs on model train sets and such. we want to run a project that is fun, and i think i can say my 4 running computers will be transfered over on or about the 16th of december till then i will probaly only have 1 running here. Nasher |
blackbird Send message Joined: 4 Nov 05 Posts: 15 Credit: 93,414 RAC: 0 |
I would welcome the idea to open Rosetta's sources - even under non-free license. Although i couldn't assist in code writing (i'm a clinical researcher), i prefer to use open source. |
andy_3_913 Send message Joined: 17 Sep 05 Posts: 9 Credit: 220,760 RAC: 0 |
"Thank you Aaron and others for this positive feedback--it is really great! This has been a lot of fun for me and the others on the team. Part of our job as scientists is to communicate what we are trying to learn to other people. for example, right now I'm in new york city (hence this response will be short!) and will soon be leaving for seminars at princeton (tomorrow), yale (tuesday), and a group of universities in nyc before returning to seattle on wednesday. I don't like travelling (I hate planes and airports!) and it is great to be able to communicate with all of you in this way instead--maybe I can skip the planes entirely soon.. Another reason we are trying to make this as fun for everybody as possible is that the single major stumbling block to progress for us is computing power--and you are making it possible for us to approach problems we never could have solved before. The problem with most biology related computation has been that the results were just not very accurate, and so computing power didn't really matter so much. The big news in our two recent Science papers (one on structure prediction, one on modelling biological macromolecules more generally), is that for the first time it is really possible to compute things like protein structures and itneractions accurately, but that it takes lots of computing power. Hence, you are all making possible what we hope will be a revolution in biology--the transformation from a purely experimental science to one in which computation plays a major role in solving problems. In other words, you are getting more feedback from us probably because we need you more! [/quote]" Sorry posting late, but what a post from a project admin; Thank the users, tell us what's going on/to happen, WOW, project admins that really seem to care about what they're doing!!!! Gobsmacked!!!!! |
Biff Send message Joined: 18 Feb 08 Posts: 1 Credit: 117,614 RAC: 0 |
Have been out of the system for some years due to ill health. Has there been any significant outcomes from the project since 2009? JW |
svincent Send message Joined: 30 Dec 05 Posts: 219 Credit: 12,120,035 RAC: 0 |
Have been out of the system for some years due to ill health. Has there been any significant outcomes from the project since 2009? I suggest looking through the thread on this message board entitled "David Baker's Rosetta@home journal" as a way of seeing the progress made in the last few years. |
Message boards :
Rosetta@home Science :
To David Baker, David Kim, Jack, & Other Admin
©2025 University of Washington
https://www.bakerlab.org