HP's NVIDIA Tesla K20 Compute Processor gives you supercomputing performance. ECC memory delivers uncompromised accuracy while an L1/L2 cache design helps improve bandwidth and data sharing. Innovative NVIDIA technologies boost application performance.
- Reach new heights in performance and system efficiency and open doors for new sets of GPU applications with Kepler GPU architectural enhancements that include the SMX streaming multiprocessor, Hyper-Q technology, and Dynamic Parallelism.
- Maximize performance and reduce data transfers from system memory with large on-board GDDR5 memory, which keeps larger data sets in local memory attached directly to the Kepler GPU.
- Enhance application data integrity with ECC memory protection, which provides critical workstation computing accuracy and reliability by protecting registry files, L1/L2 caches, shared memory, and DRAM.
- Use OpenACC or CUDA programming toolkits for C, C++ or Fortran for parallel application program development.
- Dynamic Parallelism simplifies parallel programming by enabling GPU threads to automatically spawn new threads, eliminating unnecessary program control transfers between the CPU and GPU and allowing the GPU to accelerate a broader set of algorithms.