<img height="1" width="1" style="display:none;" alt="" src="https://dc.ads.linkedin.com/collect/?pid=235929&amp;fmt=gif">


by Will Shirley, on Feb 9, 2018 9:41:37 AM

NVIDIA® CUDA® is a parallel computing platform and programming model that allows the company's Graphics Processing Units (GPUs) to provide general purpose processing.  General Purpose GPU (GPGPU) computing excels at the processing of large blocks of parallelized data.  GPGPUs break these large chunks of data down into smaller pieces, aiding computational ability by capitalizing on the high throughput design of NVIDIA GPUs.  This means real-world performance benefits to applications in many arenas where large data sets are involved, such as: signal processing, data analytics, supercomputing, finance, scientific computing and machine learning.

NVIDIA Tesla® Performance

The NVIDIA Tesla K80 GPU accelerators are the world's fastest, providing up to 2.91 TFlops of double precision and up to 8.74 TFlops of single precision performance.  Onboard Error Correcting Code GDDR5 memory--24GB onboard the Tesla K80 GPU accelerator--ensures data integrity and reliability, while data bandwidth of up to 480GB/s on the Tesla K80 GPU accelerator ensures bottlenecks do not hinder system performance.  NVIDIA GPUDirect™ technology allows additional devices on the PCIe bus (such as additional Tesla GPU Accelerators or communications interfaces) to communicate directly with each other, reducing CPU overhead load, reducing data latency.


As an NVIDIA Preferred Solution Provider, Trenton is a leader in providing solutions that capitalize on the performance benefits of GPGPU-accelerated computing.  Trenton invites System Architects and Engineers to utilize the NVIDIA GPU Test Drive to experience how your application software or one of the built-for-GPU accelerated applications, such as AMBER, NAMD, GROMACS or LAMMPS, can benefit from acceleration via a remote GPGPU cluster of NVIDIA's Tesla® K80 GPU accelerators in either Server or Workstation environments.  Obtaining a NVIDIA GPU Test Drive is simple:

  • Complete the the registration form to sign up.
  • A NVIDIA representative will contact you with instructions on utilizing a remote Tesla cluster.
  • Execute your CUDA-enabled application.
  • Determine the level of performance benefit in your particular application.

After your NVIDIA GPU Test Drive, call on Trenton's system integration expertise in deploying a Tesla-enabled Trenton Systems computing solution, such as the Trenton THS5087 HDEC Series 5U Rackmount Computer, in your application.  Contact us today.


Trenton Systems Blog

Our blogs cover the latest ruggedized computing news and company updates.


Subscribe to Updates