Project 6318 released for uniprocessor clients

By , the in Folding@Home Project - No comment
Project 6318 released for uniprocessor clients
Vincent Voelz has just put up a new project for the uniprocessor client: project 6318. This is an extension of projects 6313-6316.

The purpose of this new project is to gather additional data on the SH3 domain folding using new harmonic restraints. You can find more information on this at the previous news, where we announced projects 6313-6316.

This project is distributed by the server at IP address, and is available to uniprocessor clients with the larger units (>5 MB) option enabled. Each WU is worth 336 points, and the unit should preferably be returned within 36 days, but no later than 52 days.

Fun fact: this project outputs a pun upon the beginning of unit folding. You will see this line appear in your log:

[16:11:12] Protein: Great Red Oystrich Makes All Chemists Sane in water

Power PC OSX support to be deprecated

By , the in Folding@Home Project - No comment
There have recently been WU shortages for those running OSX on Power PC processors, due to the new cores not supporting the PPC architecture. Vijay Pande has stated that although there will still be new PPC WUs in the future, PPC OSX support for Folding@home is now in the process of being deprecated.

The overall computational power provided by the currently folding PPC processors has proven to be too small, and the effort required to support it large enough, that the Pande Group has decided that it will be more worthwhile in the long-term to reallocate the manpower currently going into PPC code maintenance, and instead consolidate it in order to better improve support for the other currently supported operating systems.

Source: Official Folding@Home blog

nVidia core 11 1.31 in auto-update on Monday, January 4th

By , the in Folding@Home Project - No comment
nVidia core 11 1.31 in auto-update on Monday, January 4th
Core 11 1.31 for nVidia GPUs has already been tested extensively and its improvements seem to be very useful for the new Project 10101. This is the reason why Stanford will be forcing an auto-update on all nVidia clients to update the core to version 1.31.

We also have some good news regarding GPU3. It looks like the work is making good progress, and a beta version may be ready in early 2010! Everything seems to be good and well in line with our estimations. ;)

Source: Official Folding@Home blog

Folding@Home in 2010

By , the in Folding@Home Project - No comment
Many people are probably asking themselves: what's in store for the project in 2010? Well, at the end of this year, we already have some answers.

Firstly, the redesign of the psummary system just ended. The new script is ready to usher in new projects and new v5 servers. This new version of the server software comes into production after the correction of a few last minute bugs. 2010 will be the year of v5 servers, new machines and new projects; good news for the stability of the entire system.

As explained in the previous news, 2010 will also see the start of the production of a new statistics system. Currently, most of the features are disabled (such as tracking the number of folded WUs project by project, for example), but downtime during updates should rapidly disappear. Folding@Home is ready to continue its expansion.

The rewrites of many parts of the v7 Folding@Home client is expected to fill many common complaints by new and existing users. We anticipate a unified client that is easy and sleek to use and set up, without the need to carry out sometimes obscure and excessive installation procedures. Vijay Pande believes that the client should be available during 2010.

The WU assignment system should also be much improved with the rewrite of the server assignment code, which has by now become very complex, often leading to quirks and downtime. This new AS (Assignment Server) code is planned for late 2010.

We are also preparing to welcome in new computing cores: GPU3, SMP2 and Protomol. GPU3 marks the transition from proprietary GPGPU technologies (CAL/Stream for ATI and CUDA for nVidia) to a unified platform: OpenCL (through the open source OpenMM program developed by Stanford). We hope that the transition to OpenCL will improve the compatibility of computing cores with new generations of GPUs without a necessitating a complete rewrite of the code. 2010 we will also show whether or not ATI will return in the OpenCL performance race with nVidia.

Protomol, which made its appearance this season, should also take up an important place during 2010. It is one of the priority projects developed in response to a focus on performance improvements, so expect to see new projects using this core next year. Protomol might also come into play as far as GPU folding is concerned, since it is designed to interface with OpenMM.

On the multi-core CPU front, 2010 will be the year of threads. The SMP2 cores (based on Gromacs and Desmond) should be launched this year, and will excise the MPI layer and processes model. We all hope that this will simplify the use of SMP and that their clients will show more equality between Windows and Linux in terms of performance. We also assume that with these new cores will come new projects, along with the ability to engage other areas of study such as on the influenza virus (primarily the 266x and 267X projects, run under SMP).

On the occasion of its 10th anniversary, the project is slowly renewing; the older tools are gradually being revamped to support the incredible growth of the project. FAH-Addict will not fail to keep you abreast of the latest developments.

Statistics system webpage optimisations

By , the in Folding@Home Project - No comment
As you may know, the statistics system is planned to migrate during January to a more efficient server to better withstand the increasing number of project users.

In preparation for this migration, the Vijay Pande's team will perform an audit of the current system to identify which parts of the system pose the most problems. During this period, you might see some bugs on the stats pages.

It seems that some users abuse the stats pages by performing more than 1,000 queries on it per day. Vijay believes that normal use should not exceed 50 queries per day; the offending IP addresses will therefore be blacklisted in order to reduce system load. If you have a script that makes a lot of queries to the system, you will need to disable it.

Remember that if you need to access any statistics using a script, you have text files available. Visit this official FAQ entry for details.

Until the new server is established, certain functions will be disabled, such as the displaying of the number of active CPUs. This database has become so overburdened that the current server can no longer efficiently perform queries on it. The statistics pages will display "N/A" for such disabled information.

All of this should be reactivated when the new system goes into production in January.

Source: Optimizations to stats web pages

New project for nVidia GPU2: p10101

By , the in Folding@Home Project - 5 Comments
In recent hours, some of you may have noticed a new WU on your nVidia GPU2 client: Project 10101.

This project is under the charge of Greg Bowman. It is distributed by a new server: It's reported to produce 548 points per WU, and the unit should preferably be returned within two days, with a final deadline being set to three days.

This project explores the lambda repressor of the villin headpiece. It is the continuation of Project 4744 (and similar) and has 1254 atoms. This project falls into the category of "large" units, so you can expect a decrease in PPD, and the operating temperature to ramp up.

We do not yet know if this project has been officially launched or if it is a test project that has mistakenly filtered through. Some people have also reported delays in downloading these units.

We will keep you informed if new information emerges.

Update as of 01:20 UTC, 29/12
Vijay Pande has posted an update about the non-crediting of points for these WUs:
"Taking a look, I see that this server isn't engaged into the stats system, so the points are safe (sitting in the log on the server), just not entered into the stats db. We've hooked it in and I reminded the researcher in charge of this machine that this wasn't done. The points should show up in the next stats update."

Update as of 12:10 UTC, 29/12
Greg Bowman has just confirmed that this project is public. It is distributed to –advmethods users.

Update as of 17:35 UTC, 30/12
This project is now distributed to everyone.

Protomol core unveiled with -advmethods

By , the in Folding@Home Project - 6 Comments
The Protomol core for uniprocessor clients has just entered the public testing phase for –advmethods flag users. It comes along with two new projects: 10000 and 10001.

This new core implements the NML (Normal Mode Langevin) method which accelerates the long-time dynamics of the proteins by a factor that can reach speeds up to a hundred times faster than those of molecular dynamics. This method searches for low frequency directives by using normal mode analysis and projects the motion of the molecule along them while resolving the nearly instantaneous motion. If you want to learn more about this method, you should read this pre-publication: Multiscale Dynamics of Macromolecules Using Normal Mode Langevin.

Based on Protomol 3.1, this core and its associated projects have the following goals:

  • To validate NML by simulating the folding and dynamics of the Fip35 WW domain.
  • To understand the role of mutations on folding.
  • To understand the activation of src Kinase, an enzyme that is involved in the onset of some kinds of cancer.

On the technical side, this core is able to take advantage of most modern CPU optimizations (SSE2, SSE3, SSSE3, SSE4.1 and SSE4.2), however, a few compatibility issues are still present on AMD processors, resulting in the core only using SSE2 on these chips. This should change in the not-too-distant future when the issues have been worked out. If you have a processor that doesn’t have the above mentioned optimizations (Pentium 3, Athlon XP, etc.), please report the behavior and the performance of this core on your machine.

For more information about the Protomol core, you should visit the Protomol official site.

The new projects are distributed by a new server ( which is located at the University of Notre Dame (Indiana) and have the following characteristics:

  • p10000 : 544 atoms, 84.48 points, preferred deadline 3.07 days, final deadline 23.04 days. This project uses conventional simulation methods.
  • p10001 : 544 atoms, 50.56 points, preferred deadline 1.84 days, final deadline 13.79 days. This project uses the NML simulation method.

Updated on 22 Dec at 22H15 UTC :
The behaviour of Project 10001 has proven to be different from that which we usually see; sometimes, the simulation enters into a state where no further computation is possible, but which is not technically a folding error. This situation usually triggers, with other cores, an Early_Unit_End or an Unstable_Machine error. With the Protomol core, these states are expected and don't trigger an error. You'll see them with a WU reported as completed (Finished unit, core status = 64) before it reaches 100%. Don't worry if this phenomenon occurs on your machine; it is perfectly normal and expected.

Happy folding!

Updates to psummary

By , the in Folding@Home Project - No comment
Joseph Coffland has created a new version of the Project Summary code, the page listing all active projects and their characteristics (deadlines, points, server, etc)

This was necessary to make the script (a near-10-year-old perl script) work with the new projects appearing in the near and distant future, and with the v5 server code. Joseph also took the opportunity to make the page compatible with the W3C XHTML standard.

You may however encounter some minor problems of missing projects or inconsistent data... these are being tracked down and fixed by Joe as he updates the system, but feel free to report any unfixed issues in the topic Joe has created on the official forums.

If your monitoring software is having problems reading the new psummary, please remember to check for updates from the developer's website.

SMP2 enters next phase of testing at Stanford!

By , the in Folding@Home Project - No comment
It has been announced on the official Folding@Home project blog that SMP2 is currently undergoing beta testing. But what is it?

SMP1 was a clever method of simulating true multithreading on multicore systems by using the MPI API in conjunction with a modified monocore (the Gromacs 3.99 core, to be exact). This method has been used since 2007, with some success, but has also shown some limits, especially on the stability side of things.

SMP2 is revolutionary, because it is a major overhaul of the Gromacs method, and uses the native capabilities of multicore CPUs. In terms of reliability, we should once again find ourselves on less treacherous ground, as the intermediary layer that is MPI - which could occasionally crash the core and/or screw up the calculations - has been removed. This core has provided a great deal of grief to Peter Kasson and the Gromacs development team.

Though... are we speaking of a single core? No; we must mention SMP2 cores. For the moment, only SMP2 A3 is being tested. This one is the heir to the SMP1 A2 MacOSX/Linux core. But other methods of calculations are also being considered and directly implemented in SMP2 in the form of the core known as Desmond.

In other important news, SMP2 also makes widespread use of the bonus system, which until recently was reserved for BigAdv units. This means the faster a unit is completed and returned to Stanford, the greater the points reward will be. We do not yet know what the basis for calculating points will be, or how performance will be positioned relative to the A1 and A2 cores today.

We will keep you informed as soon as we have more information.

Source: The Official Blog

Stats server crash of 12/11 - what really happened!

By , the in Folding@Home Project - No comment
Stats server crash of 12/11 - what really happened!
Vijay Pande, the head of the Folding@Home project, has spoken about the statistics server crash on 11 November. This failure was more serious than any previous ones incurred by the project. For the first time in the project's history, the fault has caused the loss of information between the stats server and work server. It was therefore impossible to make a points "recredit" as per usual.

The researchers have therefore opted for a "projection" in the number of points for each user, based on production recorded the day before, to get an approximation of what you really folded on that day.

Vijay Pande is perfectly aware that the points are very important in the eyes of many, despite having little value compared to the research being carried out. He is therefore thinking about upgrading the server's hardware by the end of January 2010, as the current tech is over 5 years old.

The good news is that no scientific work has been adversely affected by this problem, and the results of calculations performed during this outage are still being used by the researchers.

Source: Vijay Pande's post on the Official Forum