Skip to content

Commit

Permalink
Manually specify max_threads
Browse files Browse the repository at this point in the history
  • Loading branch information
sriharshakandala committed Jul 27, 2023
1 parent 0446116 commit 64f3bda
Showing 1 changed file with 3 additions and 3 deletions.
6 changes: 3 additions & 3 deletions src/Operators/finitedifference.jl
Original file line number Diff line number Diff line change
Expand Up @@ -3448,9 +3448,9 @@ function Base.copyto!(
Nq,
Nh,
)
kernel = @cuda launch = false copyto_stencil_kernel!(args...)
kernel_config = CUDA.launch_configuration(kernel.fun)
max_threads = kernel_config.threads
#kernel = @cuda launch = false copyto_stencil_kernel!(args...)
#kernel_config = CUDA.launch_configuration(kernel.fun)
max_threads = 256#kernel_config.threads
nitems = nnodes * Nq * Nq * Nh
nthreads = min(max_threads, nitems)
nblocks = cld(nitems, nthreads)
Expand Down

0 comments on commit 64f3bda

Please sign in to comment.