omlins / ParallelStencil.jl

Package for writing high-level code for parallel high-performance stencil computations that can be deployed on both GPUs and CPUs
BSD 3-Clause "New" or "Revised" License
312 stars 31 forks source link

CUDA 3.x GPU method overrides `CUDA.pow()` error #25

Closed luraess closed 3 years ago

luraess commented 3 years ago

The scalar_porowaves2D.jl and HydroMech2D.jl miniapp codes produce a runtime error when running on the GPU using CUDA v3.x on Julia v1.6 as the @pow() macro calling CUDA.pow() in the GPU kernels is no longer supported (CUDA.jl implements now following GPU method overrides ). See here for the full supported intrinsic list.

The solution is to replace @pow(a, b) by a^b both in the host code and the device functions (kernels).