Folding@home

Blog moving to http://folding.stanford.edu/home/blog

With our new web site, we want to unify all of our web assets under the same url and same look (as much as possible).  So, we've moved our blog to

http://folding.stanford.edu/home/blog

For now, we'll keep the old blog here, but we've moved the old posts, so eventually this site will go away as well. 

UPDATE 4: Planned server room maintenance: Monday, August 26

Looks like everything is up, except for a single server (VSP07) and its VM's associated with it.  This server is serving Core11 GPU clients, so those are off line at the moment.  Our sysadmins are working on this now.

UPDATE 3: Planned server room maintenance: Monday, August 26

The networking seems to be up, but there are a few issues.  They've got something basic going now and will resolve the remaining issues in the morning.  We are running a stats update right now.

UPDATE 2: Planned server room maintenance: Monday, August 26

Looks like the servers are up and healthy, but there is an issue with the networking which central IT is working on.  For now, we are in a holding pattern until the networking has been resolved.

UPDATE: Planned server room maintenance: Monday, August 26

The maintenance is moving along, but certain key machines are down during the move, most notably the main AS, GPU AS, and some key stats systems.  We expect the maintenance to be completed in an hour or two.

Reminder about aging cores (eg Core11, Core78)

One aspect which has dominated FAH for a decade is the continual push for new scientific approaches.  This manifests itself in terms of new scientific cores.  For example, the new GPU core (Core17) has brought huge speed improvements (especially to AMD GPUs) and involved a complete rewrite.

A negative consequence of this continuous push for improvement and progress is that newer cores often are restricted to more advanced hardware.  To help utilize as much power as is available to FAH, we tend to continue projects with older cores, but eventually, the science they can do becomes too limiting, and we must retire them.

This is the eventual fate for all cores, but is most certainly an issue sooner for certain cores, especially cores11 and 78.  While we don't have any specific end dates for either, we'd like to remind donors that those cores are reaching "end of life" status, and when they are retired, certain older hardware (eg CPUs that don't support SSE2 or older GPUs) won't be supported by FAH.  

The bottom line is that we're working to delay that as long as possible, but this post is a heads up that our support of those cores won't last a lot longer.  If I had to guess, I'd say probably within a year or so they would be retired, maybe 2 years if the existing projects need additional data.  

As always, we'll give donors more information as we know it and try to give a more specific end date when we know it.

Planned server room maintenance: Monday, August 26

We will be moving many key Folding@home servers on Monday August 26.  While much of FAH will be up, certain key systems will be down for a few hours starting in the morning (pacific time) of Monday August 26.  We'll give more updates as we move along on Monday.

A peek into Core 17 benchmarking

Our primary goal with benchmarking is "equal points for equal work."  However, making this process consistent over lots of different types of WUs and different types of hardware is tricky.  We had an internal discussion about the PPD for two projects (7810 and 8900) recently and we thought donors might find these details interesting.  

We were working to rebalance the points to make the PPD consistent, but just doing that over the wide range of hardware is difficult.  Check out the graph below which shows the PPD on the y-axis and donor GPUs sorted along the x-axis by typical PPD.  The dark line shows averages and the gray area shows error bars (variation between WUs for a given project on the same GPU type). 

What we see is that our protocol balanced the PPD on the low end, but on the high there is both bigger variation (more shaded areas) and also bigger differences on the very highest power GPUs.  In these situations, we usually go with our protocols, but this time, given all the analysis we did on it, I thought it would be interesting for donors to see these sorts of details.

It's these sorts of variations which leads to PPD fluctuations, so perhaps the main lesson here is that even with our protocols and plans, it's really hard to be consistent over all the different hardware, even when we're talking about just GPUs and just 2 projects.

New FAH web site

We've been working behind the scenes on a revamp of our web site.  It went live today (http://folding.stanford.edu/home).  This is part of our larger plan to make FAH more friendly and easy to use, especially to non-experts.  

With that said, we're now thinking about next steps to make FAH more fun and appealing to experts, such as computer enthusiasts and gamers.  We're in the early stages of deciding what would be useful there.  If you have ideas, please do give us some feedback on our forum at this thread:

http://foldingforum.org/viewtopic.php?f=16&t=24532

Welcome to FahCore 17!

We are proud to announce that our latest GPU core, FahCore 17, was recently moved from beta to advanced testing, the last quality assurance step before a full release. As we previously mentioned, this core is a significant step for us. FahCore 17 is a complete overhaul from our previous GPU cores. It brings a cleaner and more streamlined codebase, new serialization mechanisms that allow us to set up diverse simulations, and improved stability. Its use of OpenCL has united our development, allowing the single core to run on both Nvidia and AMD cards, and theoretically any OpenCL-capable device. It is also our first GPU core to run natively in Linux, although we are only supporting Nvidia GPUs there for the time being as we wait for AMD's Linux drivers to mature a bit more. Overall, this core sets a strong foundation for the future of GPU core development.

On AMD cards, FahCore 17 is about 10 times faster than the old GPU cores, and on Nvidia it's about twice as fast. This is mainly due to its OpenMM 5.1 base, which contains many optimizations which deliver a significant speedup. One optimization in particular that we are waiting for is CUDA JIT, a just-in-time compiler that Nvidia may be introducing into its drivers in the coming future. Not only will this technology allow us to offer support for the CUDA platform with FahCore 17, but the JIT compiler is likely to deliver a massive speedup. For the time being, we continue to work at finding additional optimizations on our end. We have also successfully tested FahCore 17 with extremely large proteins (500,000+ atoms), which are on par with the ones used by "bigadv" CPU projects.

To run FahCore 17, you need a Fermi GPU or better and Windows or Linux, or a AMD HD5000 or better and Windows. It also currently requires proprietary drivers from these vendors. You can test FahCore 17 by adding the "client-type = advanced" setting into the extra core options in the V7 client, as in the Configuration FAQ. Another excellent resource is the GPU FAQ which describes why GPUs are so helpful to us.

We'd like to thank all the alpha testers on FreeNode's #fah IRC channel, as well as the beta testers on foldingforum.org, who have all helped us bring the core to this point!