Folding Forum

FAHbench v1.2.0 uses old OpenMM 5.1 which is not used anymore in FAH GPU cores. Use FAHbench v2.3.2 using OpenMM 6.3 which matches current FAHcore_21 https://github.com/fahbench/fahbench/releases I know, but the old-version benchmark still clearly shows the difference between the various devices. I...

FAH v1.2.0 link
https://dropmefiles.com/e8OwO

ok, this my RTX2060
OpenCL SP
Explicit:
113.341 ns/day
Implicit:
603.647 ns/day

ok, this my RTX2060
OpenCL SP
Explicit:
113.341 ns/day
Implicit:
603.647 ns/day

OpenCL SP


Double precision in FAH is pointless?

OpenCL double precision