XStream® GPU Acceleration
Built-in EM Simulation Acceleration via Graphics Processing Units
XStream tremendously improves EM simulation performance by leveraging the powerful NVIDIA graphics processing units (GPUs) available in modern video cards to make ultra-fast FDTD numerical computations. Using NVIDIA’s latest generation GPUs, XStream enables XF calculations to finish in minutes as compared to hours or even days using a CPU only.
XStream requires at least one CUDA capable GPU. Remcom recommends NVIDIA CUDA-enabled Fermi or Kepler series GPUs. For specific details, please visit NVIDIA's website.
MPI + XStream for GPU Clusters:
Exceptional performance is possible with XStream alone but is limited to the available resources of a single computer. Combining XStream with the power of MPI (Message Passing Interface) allows multiple computer nodes to work together using the memory and computing power of all GPUs in the group, massively increasing processing speed and available memory. Whether you want to pool existing resources for cost savings or accelerate problems you never thought possible, MPI + XStream provides you with the market’s leading technology. Learn more about MPI + XStream…
Custom Configured Hardware:
Remcom’s partners can provide customers with custom XFdtd systems outfitted with the latest in GPU technology. These Remcom-certified systems optimize XFdtd’s performance by building the recommended GPUs directly into the product.
While a custom-engineered system is not necessary for using XStream, these offerings deliver cluster level computing performance for those users that prefer additional power.
|XFdtd Certified GPU Systems by Exxact
Performance Study: XF7's MPI + GPU Technology and Hardware Comparisons
With XFdtd EM Simulation Software, there is no limit to the resources you can exploit to solve your EM calculations. This report quantifies the performance profile of XFdtd’s GPU and MPI technologies. We demonstrate the pros and cons of different combinations of equipment and techniques, including cost considerations for those researching available hardware. Download the free paper.
A license to employ a single GPU for EM computation is included in both XF7 Pro and XF7 Bio-Pro, with additional capability for multiple GPUs available as a purchased option. Contact Remcom for details.
NVIDIA and CUDA are trademarks and/or registered trademarks of NVIDIA Corporation in the United States and other countries.