GE GRA112 Graphics Board パンフレット

gp gpu  

•  

www.ge-ip.com/gp gpu

2

The world of high-performance computing 
is undergoing a revolution, thanks to 
advances in General Purpose computing 
on Graphics Processing Units (GPGPU). The 
idea behind GPGPU is to use a GPU, which 
typically handles computation for computer 
graphics only, to perform parallel computa-
tion in applications that have traditionally 
been handled by the CPU. 

A multi-GPU platform hosted by one or 
more CPUs is able to perform heteroge-
neous computing, harnessing the parallel 
computing power of the many-core GPUs to 
provide very large increases in performance 
with minimal programming complexity.

Additionally, programmers are helped with 
software development environments such 
as Compute Unified Device Architecture 
(CUDA) and OpenCL, which allow them to 
harness the many-core, parallel processing 
capabilities of the GPGPU platforms. 

While greatly increasing functional capa-
bility, the GPGPU platform also delivers the 
performance with far less size, weight and 
power (SWaP). This results in significant 
savings in cost, risk, and time-to-market. 

Now these benefits are fully available for 
rugged military and aerospace applications. 
With the introduction of a full range of GE 
Intelligent Platforms rugged GPGPU boards 
and systems, the advantages of GPGPU are 
no longer confined to controlled environments 
at universities, research centers and hospitals. 

The unique partnership between GE Intel-
ligent Platforms and NVIDIA allows for new 
product development using NVIDIA GPUs 
based on the award-winning CUDA architec-
ture, for Military and Aerospace applications.  

2 cores

96 cores

Rugged GPGPU COTS Boards 

for Military and Aerospace

GPGPU platforms deliver new levels of  
performance for size, weight and power 
(SWaP) constrained mission payloads.

Lab-proven technologies rugge-
dized for the rough and tumble 
world of military applications