| | 1 | '''New in CUDA 8''' |
| | 2 | |
| | 3 | '''Pascal Architecture Support''' \\ |
| | 4 | * Enhance performance out-of-the-box on Pascal GPUs |
| | 5 | * Simplify programming using Unified Memory including support for large datasets, concurrent data access and atomics |
| | 6 | * Optimize Unified Memory performance using new data migration APIs |
| | 7 | * Increase throughput at ultra-fast speeds using NVIDIA® NVLINK™, new high-speed interconnect |
| | 8 | '''Developer Tools''' \\ |
| | 9 | * Identify latent system-level bottlenecks using critical path analysis |
| | 10 | * Improve productivity by up to 2x with faster NVCC compile times |
| | 11 | * Tune OpenACC applications and overall host code using new profiling extensions |
| | 12 | '''Libraries''' |
| | 13 | * Accelerate graph analytics algorithms with nvGRAPH |
| | 14 | * Speed-up Deep Learning applications using native support for FP16 and INT8, support for batch operation in cuBLAS |
