Skip to content

Use grid-stride loop for fill! #691

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
christiangnrd wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Use grid-stride loop for fill! #691

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 12 additions & 2 deletions src/host/construction.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,12 +14,22 @@ function Base.fill!(A::AnyGPUArray{T}, x) where T

@kernel function fill_kernel!(a, val)
idx = @index(Global, Linear)
@inbounds a[idx] = val
stride = prod(@ndrange())
while idx <= length(a)
@inbounds a[idx] = val
idx += stride
end
end

# ndims check for 0D support
kernel = fill_kernel!(get_backend(A))
kernel(A, x; ndrange = length(A))

# Calculate ndrange to ensure that a total grid size >typemax(UInt32) is never
# chosen. Grid stride to accomodate grid size limitations on AMD and Metal backends
len = length(A)
ndrange = cld(len, cld(len, typemax(UInt32) - 1024))

kernel(A, x; ndrange)
A
end

Expand Down
Loading