mirror of
https://github.com/NVIDIA/cuda-samples.git
synced 2026-01-10 12:37:49 +08:00
f288b2e261
Update kernel loop bounds handling, main loop data copy to avoid incorrect reuse of output results. --------- Authored-by: Francesco Rizzi <francesco.rizzi@ng-analytics.com>
6. Performance
alignedTypes
A simple test, showing huge access speed gap between aligned and misaligned structures. It measures per-element copy throughput for aligned and misaligned structures on big chunks of data.
transpose
This sample demonstrates Matrix Transpose. Different performance are shown to achieve high performance.
UnifiedMemoryPerf
This sample demonstrates the performance comparision using matrix multiplication kernel of Unified Memory with/without hints and other types of memory like zero copy buffers, pageable, pagelocked memory performing synchronous and Asynchronous transfers on a single GPU.
cudaGraphsPerfScaling
This sample demonstrates the performance characteristics of cuda graphs. It is focused on how the apis scale with graph size.