Jacket MGLMulti-GPU Option for a Single Node with 2 to 8 GPUs |
Documentation |
|
Jacket MGL enables MATLAB® code to run on 2 to 8 GPUs in a single computer. There are two modes of operation with Jacket MGL, as follows:
Executing code, such as FOR-loops or monte carlo simulations, with Jacket MGL can dramatically accelerate time to solution on workstations and personal supercomputers. Example Jacket MGL Usagematlabpool 4; % create 4 PCT workers %% Jacket assigns each PCT worker to a GPU parfor n = 1:10000, gpu_var = gdouble( ... ); % offload to GPU ... = conv2( gpu_var, ... ); % GPU convolution ... % continue GPU and CPU computations end |
 |
Licensing DetailsThe Jacket base license includes support for one (1) GPU only. To purchase support for additional GPUs, upgrade your license with the MGL addon. The MGL addon increases the number of supported GPUs on a single designated computer. As an example, if you want to support two (2) GPUs on a single workstation, you would purchase one (1) Jacket Base License and one (1) MGL Addon. As another example, if you want to support four (4) GPUs, you would purchase one (1) Jacket Base License and three (3) MGL Addons. |
|
Jacket HPCMulti-GPU Option for Multiple Nodes with 8+ GPUs |
Documentation |
Jacket HPC enables MATLAB® code to run on 8 or more GPUs in a cluster, server, or cloud configuration. When a single host is not capable of driving more GPUs, simply add GPUs to another host on the network and Jacket HPC will take care of the rest. CPU clusters may be upgraded through the installation of GPUs, significantly increasing the cluster's computational capability without investing in new development for specialized GPU code.
Jacket HPC is targeted at the GPU cluster market. Distributed clusters of GPUs, that contain more than eight (8) GPUs, are well suited for Jacket HPC.
Jacket HPC requires the Parallel Computing Toolbox™ (PCT) and the MATLAB® Distributing Computing Server™ (MDCS). With the addition of parallel constructs, such as PARFOR and SPMD, pre-existing code may be dispatched across all GPUs and CPUs in a cluster, server, or cloud. No additional code revision is necessary.
Example Jacket HPC Usagematlabpool 4; % create 4 MDCS workers %% Jacket assigns each MDCS worker to a GPU parfor n = 1:10000, gpu_var = gdouble( ... ); % offload to GPU ... = fft( gpu_var, ... ); % GPU FFT ... % continue GPU and CPU computations end Executing large scale codes with Jacket HPC and GPU clusters can dramatically accelerate time to solution while minimizing the programming time associated with leveraging these resources. |
|
Licensing DetailsTo request a price quote for the Jacket HPC license addon, contact us. Jacket HPC is not available for purchase online. |
|