StorySoftware
NVIDIA introduces CUDA for general-purpose GPU computing
NVIDIA introduced CUDA, a way to program its graphics chips in C for general computing, launched with the GeForce 8800 and its 128 processor cores. NVIDIA said the parallel cores could solve some problems up to 100 times faster than traditional approaches.
- CUDA launched on GeForce 8800 with 128 processor cores for parallel computation
- Enables hundreds of cores to communicate, synchronise and share data via Parallel Data Cache
- Claimed up to 100x speedup for computation-intensive applications in product design, engineering and data analysis