Jacket (software)

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Developer(s) AccelerEyes
Last release 2.3 / October 22, 2012 (2012-10-22)
Written in C, CUDA
Operating system Cross-platform[which?][1]
Type Technical computing
License Proprietary
Website accelereyes.com/products/jacket

Jacket is a numerical computing platform enabling GPU acceleration of MATLAB-based codes. Developed by AccelerEyes, Jacket allows GPU-based matrix manipulations, plotting of functions and data, implementation of algorithms, and interfacing with programs written in other languages, including C, C++, CUDA, and OpenCL.


Jacket was created in 2007 by Gallagher Pryor, co-founder of AccelerEyes. Other original inventors and co-founders include James Malcolm, Tauseef Rehman, and John Melonakos.[2] Jacket was designed to enable high-level programming of GPU devices for rapid prototyping work.

In an October 2008 article,[3] the number of current Jacket users is said to be in the thousands, and includes engineers and scientists in technology companies such as GE, Google, Toyota, and Morgan Stanley.

Jacket sales were discontinued on 12 December 2012. Existing functionality will be integrated into MATLAB's parallel computing toolbox.[4]

Jacket use[edit]

Jacket is built around the M (or MATLAB) language. Jacket code is executed at the MATLAB command prompt, >> , in the Command Window, or as a .M script.[5] Jacket introduces new GPU-specific data types into MATLAB. Once a GPU data structure has been created, any operations on that GPU matrix are performed on the GPU rather than the CPU. To turn off GPU computation, simply cast the data back to the CPU using one of the MATLAB data types, e.g. double. These functions are used as follows:

>> GPU_matrix = gdouble( CPU_matrix );
>> GPU_matrix = fft( GPU_matrix );
>> CPU_matrix = double( GPU_matrix );

Functional coverage[edit]

Jacket includes GPU versions of most MATLAB functions[6]

Multi-GPU support[edit]

The Jacket Multi-GPU License (MGL) extends the single GPU support of the base Jacket product to as many as eight (8) GPUs in a single system image machine.[7] Where high-performance computing (HPC) resources, such as GPU clusters, are already in place, Jacket HPC can be leveraged to run M code across the clusters or Cloud services with 8 GPUs or more.[8]

Graphics support[edit]

Jacket includes a Graphics Library for coupling of GPU visualizations with GPU computation.[9] GPU versions of common MATLAB visualization commands are available, including gplot, gsurf, gscatter3, etc.[10]

See also[edit]


External links[edit]