Folding@home
Folding @ home (often short F @ H or FAH ) is a project of Stanford University to simulate the folding of proteins.
The project
Folding means (protein folding ) takes an amino acid sequence necessary for a protein function spatial structure. Mis-folding ( misfolding ) are discussed (Alzheimer, BSE and Creutzfeldt- Jakob disease, or cancer ) in the context of disease development. The aim of the project is by distributed computing to understand the spatial structure or the structure of proteins and thus for understanding the formation and healing of diseases that result.
If the protein folding only simulated on the computers at Stanford University, this would take several decades, despite the high computing power of the university computer. The aim of Folding @ home, therefore, is to distribute the computing power to as many other computers (Distributed Computing). The computers involved are operated on a voluntary basis by companies, public bodies and private individuals who place their computers and computing capacity of the project free of charge. Thus, the computational capacity is increased by a multiple, because the host only needs to process the results ready created. Currently, the project has 162 983 active CPUs ( with a total of 291 397 computing cores ) and 68 373 active GPUs and provides a computing power of 42.875 or 20.693 petaflops x86 native petaflops ( at 1 April 2014). Compared with the currently fastest supercomputers worldwide Folding @ home so would the first place (as of April 2014) are taking. Currently, several projects are underway to accelerate the computing power by a factor of 100, without the user would need for new hardware.
The software
Every user on a PC with Windows, Mac OS X or Linux can download a program that operates as a service in the background. The current version ( 7.4.4 ) supports single-core and multi-core processors as well as Windows and graphics cards from Nvidia and AMD. For the PS3, there was also a version that was set but now. With over one million PS3 - participants of this project is as yet most powerful distributed computing network in the Guinness Book of World Records.
For notebooks, there is the possibility that F @ H rests with his work, when that gets its energy from the batteries, which extends the run time.
Points
As with many other projects that use distributed computing, should be prepared in Folding @ home statistics in the form of a points system over the contributed computing power. Average 500-550 projects run on Folding @ home, where there are individual basis points for each project, which are set by a reference PC. For each unit of work has already been calculated ( Work Unit ) of a project, the user receives the appropriate basis points. These points, however, can multiply (bonus points), a Work Unit will be completed faster. Prerequisite for obtaining points is a registration with the project with a user name. To receive bonus points, in addition to the use of a so-called " Passkey " is required.
Each user can choose whether its computing power anonymous (no points ) is only counted under his user name or for a team.
GPU support
The Folding @ home client can be used for the calculation, depending on the settings for the CPU and the GPU use. Supported graphics cards from Nvidia and AMD. Currently, however, the GPU support is available only on Windows. Prerequisite for Nvidia graphics cards is the CUDA technology (from G80 with GeForce driver from 174.55 ). AMD graphics cards are supported from the HD5000er series. The graphic units of all APUs by AMD, whether desktop or laptop, can now also be used. Here, the V7 client falls back to the default OpenCL.
Results
Total so far 114 publications were published as a direct result of Folding @ home. Some data are available to all free at leisure.
Related Projects
Rosetta @ home, Predictor @ home and POEM @ home were or are projects that have the same goal, but use other methods.