Folding@home

Comparison between FAH and Anton's approaches

Right now, the two most powerful supercomputers for studying protein folding are Folding@home and a very impressive special purpose computer from DE Shaw Researched, called ANTON.  We're often been asked "how do they compare?"  The approaches are very different, so comparisons aren't completely straightforward.  ANTON takes the traditional approach to studying protein folding, where one performs a few (often 1 or 2) long trajectories to study the process.  Folding@home takes a statistical approach, which has two primary benefits:  1) it can access folding on dramatically longer timescales (milliseconds, instead of microsecond folding events over a single long trajectory) and 2) it can give statistically significant results on those long timescales.

The main concern about the method in FAH is that since it is such a radically new approach, does it work reliably?  Previous tests of FAH have been to experiment, which is the gold standard test, but also brings in other issues, such as how good are our models of reality.  Thus, while FAH's approach has done well compared to experiment, it is useful to compare FAH and ANTON directly, since they use the same models, etc.  Comparison of our statistical approach (using Markov State Models, aka MSMs) directly with data from ANTON would go a long way to showing that the MSM approach works for even non-trivial systems (they have been previously tested for long dynamics on small systems).

In a recently published paper, we make this comparison.  By applying MSMs to data from ANTON, we find that FAH's approach (MSMs) can reproduce the long timescales in ANTON data very well.  Moreover, we also find that the MSM approach can find important new features missing in the more traditional analysis approach originally applied to the ANTON data, relevant for understanding folding and function.  

For us, this is exciting since it shows the capabilities of the MSM method.  However, I want to stress that perhaps the most exciting part is how ANTON and FAH could be used together.  A run on ANTON followed by more thorough sampling in Folding@home could be the best of both worlds.

 

PS Here's the abstract for our paper (http://pubs.acs.org/doi/abs/10.1021/ja207470h?prevSearch=lane%2Bpande&searchHistoryKey=):

Two strategies have been recently employed to push molecular simulation to long, biologically relevant timescales: projection-based analysis of results from specialized hardware producing a small number of ultra-long trajectories and  the statistical interpretation of massive parallel sampling performed with Markov state models (MSMs). Here, we assess the MSM as an analysis method by constructing a Markov model from ultra-long trajectories, specifically two previously reported 100 µs trajectories of the FiP35 WW domain (Shaw et. al. (2010) Science, 330: 341-346). We find that the MSM approach yields novel insights. It discovers new statistically significant folding pathways, in which either beta-hairpin of the WW domain can form first. The rates of this process approach experimental values in a direct quantitative comparison (timescales of 5.0 µs and 100 ns), within a factor of ~2. Finally, the hub-like topology of the MSM and identification of a  holo conformation predicts how WW domains may function through a conformational selection mechanism

Update on the v7 client – v7.1.38 is now released

We've released the latest v7 client (7.1.38) in our forum.  I've pasted the update from Joseph Coffland (lead programmer), outlining the progress so far.

Documentation
Installation and user guides can be found here:
FAHControl -> The Graphical User Interface (GUI) what controls the Slots.
FAHViewer -> It shows the protein being folded, if applicable.
Pictorial Installation Guide (Windows) -> A detailed pictorial guide on the V7 installation.
Installation Guide (Windows) -> A brief guide on Windows installation.
Installation Guide (Linux) -> A guide for Linux installation.
Installation Guide (OSX) -> A guide for OSX installation that is in progress.
Client Remote Interface -> Documentation for 3rd party developers.
Main Page -> Main page of the V7.

Getting Help
Aside from the documentation the best place to get help is in this forum. If you do have a problem post a message. There are many knowledgeable people ready and willing to help. Keep in mind, we greatly appreciate thorough reports delivered by patient people who can keep a cool head even when things go wrong.

Bugs/Tickets
Open Tickets Ordered by Milestone and Priority
Active Tickets by Change Time
Note: Some tickets may be closed because they are fixed in an upcoming alpha release but are not yet fixed in the beta release.

_______________________________________________________________________________________

 

Beta Testers,

I've been focusing on addressing the issues which most affect v7 client users and which may help those still running v6 clients to make the decision to upgrade to v7. Your feedback has been very important, not only in finding and solving bugs but in prioritizing development time. One donor had these concerns with using v7.1.33:

Jesse_V wrote:I'd be happy to try out v7.1.33 and its upgrades if I was confident that my participation wouldn't hurt science. I am concerned with a number of bugs in v7, things such as FAHClient taking up way too much CPU time and slowing things down. I also am concerned about v7 not being able to talk to the servers right, especially when errors occur.

Although the issues Jesse_V mentioned only affected a small portion of the v7 users they were legitimate concerns. I made sure these issues were thoroughly addressed in this release. Concerns like this are exactly what we need to hear about.

If you are also concerned about v7 affecting the science of Folding@home I have this to say. v7 is more likely to increase your contribution but if you do have a problem, and you report it, you are potentially helping thousands of users donate computing time to Folding@home so it's more likely a net gain. However, beta testers often have to be satisfied with peer recognition and praise rather than points.

If you've been on the fence about v7 now is a great time to upgrade. Here are some of the other comments we've been getting:

jimerickson wrote:i am running windows 7 ultimate 64bit and windows 8 developer preview release. everything went smoothly. . .keep up the excellent work.

 

DarkFoss wrote:The upgrade from v7.1.24 to 7.1.33 went flawlessly. . .install took less than a minute and it resumed folding the wu's they finished without issue. . .

 

OEOTS wrote:Congrats on the new release guys. with V7 you've made it all so much easier.

 

soya_crack wrote:Wow, that is one of a big advance. That was what F@H just needed. It's awesome. Easy to understand and just kickin' ass, the slot thing is what I was missing from BOINC.

 

WhiteLion wrote:Awesome! :) The best client i ever use, perfect for newbies, easy setup. . .

There are still several issues that need to be addressed before we go to the front page with the v7 client. Specifically the tickets listed in Milestone Open Beta Phase 2 under ticket report 3. Log filtering (#157), better ETA calculation (#395) and better handling of cores that don't terminate cleanly (#563) are the top three items.

There is more work to do but we are getting ever closer to a front page release of the v7 client. Thanks for all your bug reports and feedback. Keep telling us what issues are most important to you and we will do our best to address them.

Joseph Coffland
Folding@home Developer
Cauldron Development LLC

Documentation
Installation and user guides can be found here:
FAHControl -> The Graphical User Interface (GUI) what controls the Slots.
FAHViewer -> It shows the protein being folded, if applicable.
Pictorial Installation Guide (Windows) -> A detailed pictorial guide on the V7 installation.
Installation Guide (Windows) -> A brief guide on Windows installation.
Installation Guide (Linux) -> A guide for Linux installation.
Installation Guide (OSX) -> A guide for OSX installation that is in progress.
Client Remote Interface -> Documentation for 3rd party developers.
Main Page -> Main page of the V7.

Getting Help
Aside from the documentation the best place to get help is in this forum. If you do have a problem post a message. There are many knowledgeable people ready and willing to help. Keep in mind, we greatly appreciate thorough reports delivered by patient people who can keep a cool head even when things go wrong.

Bugs/Tickets
Open Tickets Ordered by Milestone and Priority
Active Tickets by Change Time
Note: Some tickets may be closed because they are fixed in an upcoming alpha release but are not yet fixed in the beta release.

Here are the change logs:

v7.1.38:

• Fixed network connection dropping.

v7.1.37:

• Added missing wraplabel.py file to FAHControl.
• Changed socket error message verbosity.
• Fail WU on UNSTABLE_MACHINE immediately & return for partial credit. #615

v7.1.36:

• Fixed a potential socket connection bug. Maybe related to #734.
• Added several NVidia cards to GPUs.txt. #737.
• Improved Linux on battery detection. #738.
• Print WU error state on WU status line.
• Emit correct exception on FAH transaction failure. #615.
• Fixed debian package install core permissions problem. #732.
• Removed core byte order warning. #602.
• Added GPL link to FAHControl about. #736.
• Ask user, team, passkey and mode during .deb package install. #739.

v7.1.35:

• Added 'Enchanter' theme. #731
• Renamed 'Wimp' to 'Windows-Default'. #731
• Unminimize FAHControl window on unhide. #567
• Better core download failure message. #161
• Cleaned up project descriptions using html2text.py.
• Store project data in client DB.
• Use system default font size. #733
• Added project info to viewer. #575.
• Added clickable buttons to viewer.
• Fixed FAHViewer crash introduced in v7.1.34.
• Fixed mouse wheel scrolling in FAHControl. #463.
• Fixed color difference for text boxes. #698.
• Changed FAHControl window name. #711.

v7.1.34:

• Fixed CPU consumption in client connections. #702
• Really fixed "Wrong architecture" bug on 32-bit Ubuntu. #599
• Only warn on config errors. #722
• Log error and continue of command server fails to initialize.
• Fixed Slot configuration text. #717
• Use -1 or 0 for CPUs default to be consistent with GPU options. #717
• Disabled no longer supported AMD X1300 - 1900 GPUs.
• Added "OpenGL Render" to info in FAHViewer. (For blacklisting)
• Added 'override-blacklist' option to FAHViewer. (Nothing black listed yet)
• 'OK' -> 'Save' in FAHViewer preferences window. #724
• Fixed NVIDIA_DEV.1244.01 = "NVIDIA GeForce GTX 550 Ti" detection.
• Added the 'Wimp' theme and win32 theme engines. #723
• Made 'Wimp' theme the default in Windows. #713
• Added heartbeat to viewer<->client connection to timeouts dead connections.
• Stop trying FAILED, FAULTY and DUMP reports if WS connection was made. #728
• Check WS server versions for unreasonable values. #728.

Here are the highlights and some of the technical details of the changes in this open-beta:

Improved network handling:
A number of networking related issues were solved. The run away CPU usage problem was caused by incorrectly handling dropped connections. Periodic connection loss on Windows machines was caused by not correctly handing a quirk in the Windows sockets API which was the cause of the greyed out FAHControl that some people were experiencing. Also, dropped connections are now being cleared out of the client much more quickly. This keeps dead connections from accumulating which caused problems in a few cases.

Improved communication with older WS:
Several problems with communication with older WS were fixed. This version does a better job of detecting older WS and aborts the error loops that the previous client got into when incorrectly trying to upload DUMP reports to the old servers. This was the cause of the repeated upload errors many people saw in their logs. Also, a bug was fixed in the handling of UNSTABLE_MACHINE core return code which was the start of many of these DUMP communication loops.

GPU detection:
More cards were added to the whitelist and some older, no longer supported, GPUs were removed.

Debian package improvements:
A permissions problem was fixed that was causing WU loss after reinstall. The installer also now asks for user, team and passkey if /etc/fahclient/config.xml does not already exist. RPMs and OSX packages still need this feature.

Added a native Windows theme (Wimp):
The Wimp (Windows Implementation) theme was added to FAHControl. This theme gives a true native look on all Windows platforms because the theme engine actually uses the Windows system to render the GUI components. This is now the default theme for Windows platforms. Still need something like this for OSX.