Dear All,

Would like to ask for your help withe a question related to the performance of NAMD 2.14b1 (NAMD_2.14b1_Win64-multicore-CUDA package) on Windows 10, which is run on a single Intel® Core™ i9-9900K Processor (16 threads) equipped with GPU NVIDIA Quadro RTX 4000 (Driver version 26.21.14.4292) [R440 U8 (442.92)]. CUDA Toolkit v10.2.89.

I launch NAMD with the command ./namd2 +p16 +setcpuaffinity , but I only get a very small usage of both CPU (~ 12%) and GPU (~15%).

It seems that +p16 parameter has no effect (I can change it to any number without any effect on the performance).

Would greatly appreciate your help,

Best regards,

More Stas Engel's questions See All
Similar questions and discussions