NVidia releases CUDA 5
Oct 17, 2012 by CGP Staff
0
|
NVidia has released a new version of its parallel computing platform and programming model for accelerating applications on GPUs. Highlights include:
- Dynamic Parallelism – Brings GPU acceleration to new algorithms
- GPU-Callable Libraries – Enables third-party ecosystem
- GPUDirect Support for RDMA – Minimizes system memory bottlenecks
- Nsight Eclipse Edition – Generate CUDA code quickly and easily