The Stochastic Simulation Algorithm on the Cell Broadband Engine
From PBDN
Contents |
Background
The regulatory networks in a biological cell are very complicated and not very well understood. A better understanding of them is possible using mathematical and computational models in combination with experiments. This is the goal in the emerging field of systems biology. The standard models are based on the reaction rate equations of the chemical reactions in the cell. These models work well when the concentrations of the molecular species are high, but this is seldom the case in a cell. Instead, a stochastic model for the number of molecules of each species is much more accurate. If the system is not well-stirred, then the spatial distribution of the species is also important, leading to very high dimensional problems.
The standard method to simulate these systems is to use the Stochastic Simulation Algorithm (SSA) by Gillespie[1] . It generates a statistically correct time trajectory of the system. There is a modified algorithm for problems with spatially varying species[2] . The purpose of this project is to implement SSA and the version of it suitable for space-dependent problems on a novel computer architecture and apply it to problems of biological interest.
The Cell Broadband Engine
The Cell Broadband Engine is a heterogeneous multicore microprocessor developed by IBM, Toshiba, and Sony Computer Entertainment to power the Sony PlayStation 3 games console.[3]
In summary, it consists of one conventional "master" processor, the PPE, and eight "slave" vector processors, the SPEs, each with 256kB of local SRAM, on a single die.
For "embarassingly parallel" applications in scientific computing, such as Markov-chain Monte-Carlo methods, with data sets sufficiently small that they fit within the 256kB local store of the SPE, Cell BE processors should achieve performance proportional to the number of available SPEs, even without taking advantage of the vector-processing capabilities of the SPE.
Timeline
A preliminary project plan.
Student
Advisor
Reviewer
References
- ↑ Gillespie, Daniel T. (1976), "A general method for numerically simulating the stochastic time evolution of coupled chemical reactions", J. Comput. Phys. 22 (4): 403-434
- ↑ Hattne, Johan; Fange, David & Elf, Johan (2005), "Stochastic reaction-diffusion system with MesoRD", Bioinformatics 21 (12): 2923-2924, DOI 10.1093/bioinformatics/bti431
- ↑ IBM Corp.; Toshiba Corp. & Sony Computer Entertainment Inc. (2007-10-11), Cell Broadband Engine Architecture, IBM Systems and Technology Group, <http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/1AEEE1270EA2776387257060006E61BA/$file/CBEA_v1.02_11Oct2007_pub.pdf>
Back to The Stochastic Simulation Algorithm on the Cell Broadband Engine main page.