Microway HPC Clusters and
Servers seamlessly integrate the NVIDIA® Tesla
K10 GPU Computing Accelerators. Solve your most important HPC challenges with
a massively multi-threaded, parallel-computing architecture, and the CUDA™
C, C++, and Fortran language environment and tool suites.
The lastest Tesla K10 GPUs
provide the most cutting edge single-precision performance
and must-have HPC features. Solutions are available for all scales and budgets.
Tesla K10 GPU Computing Accelerator highlights
Up to 3072 floating point processor cores
Each of the 2 GPU chips on a K10 GPU Accelerator provides 1536 cores
SMX (streaming multiprocessor) design delivers up to 3x more performance per watt compared to the Fermi architecture
Tesla GPU's CUDA cores are designed from the ground up for HPC
Over 4.5 TFLOPS Single and up to 190 GFLOPS Double Precision Floating Point throughput
IEEE 754 single & double precision floating point units
Up to 8GB high speed GDDR5 memory with ECC function*
External DRAM is ECC protected
384-bit memory interface with 320 GB/sec memory bandwidth (160 per GPU)
Accessible programming environments
Native NVIDIA CUDA C/C++ environment
PGI Accelerator Fortran for CUDA GPUs available
OpenCL and DirectCompute capable
Fits seamlessly into HPC environments
K10 series GPUs may be integrated into specially designed GPU servers and clusters that provide the necessary cooling
*With ECC enabled, memory available to the user
will be decreased by 12.5%. 8G raw becomes 7G usable with ECC
Harness Microway's GPGPU Expertise for Your HPC Needs
NVIDIA Tesla Preferred Partner
Legendary Service and Technical Support
Full Software Integration and Flexible Offerings:
Tesla K10 GPU Accelerator
Tesla Fermi C2075, M2075, M2090
Tesla Fermi S2075, S2090 4-GPU Servers
Microway OpenCL™ Toolkit: SimpleCL™ for easy cross-platform OpenCL development
PrivacyLegal
Our mission is to provide customers with leading edge technologies for high performance computing solutions (HPC). We establish and maintain industry recognized products and expertise for beowolf cluster interconnect (InfiniBand), clustre management and HPC storage solutions.
