Skip to content
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

add field_matrix_solve! reproducer #1750

Merged
merged 2 commits into from
May 22, 2024
Merged

add field_matrix_solve! reproducer #1750

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
cd8d2d7
add field_matrix_solve! reproducer
juliasloan25 May 22, 2024
50ae409
Bump patch. Fix single_field_solve for non-extruded spaces, update news.
charleskawczynski May 22, 2024
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
10 changes: 10 additions & 0 deletions NEWS.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,16 @@ ClimaCore.jl Release Notes
main
-------

v0.14.5
-------

- ![][badge-🐛bugfix] `run_field_matrix_solver!` was fixed for column spaces, and tests were added to ensure it doesn't break in the future.
PR [#1750](https://github.com/CliMA/ClimaCore.jl/pull/1750)
- ![][badge-🚀performance] We're now using local memory (MArrays) in the `band_matrix_solve!`, which has improved performance. PR [#1735](https://github.com/CliMA/ClimaCore.jl/pull/1735).
- ![][badge-🚀performance] We've specialized some cases in `run_field_matrix_solver!`, which results in more efficient kernels being launched. PR [#1732](https://github.com/CliMA/ClimaCore.jl/pull/1732).
- ![][badge-🚀performance] We've reduced memory reads in the `band_matrix_solve!` for tridiagonal systems, improving its performance. PR [#1731](https://github.com/CliMA/ClimaCore.jl/pull/1731).
- ![][badge-🚀performance] We've added NVTX annotations in ClimaCore functions, so that we have a more granular trace of performance. PRs [#1726](https://github.com/CliMA/ClimaCore.jl/pull/1726), [#1723](https://github.com/CliMA/ClimaCore.jl/pull/1723).

v0.14.0
-------

Expand Down
2 changes: 1 addition & 1 deletion Project.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
name = "ClimaCore"
uuid = "d414da3d-4745-48bb-8d80-42e94e092884"
authors = ["CliMA Contributors <[email protected]>"]
version = "0.14.4"
version = "0.14.5"

[deps]
Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
Expand Down
30 changes: 21 additions & 9 deletions src/MatrixFields/single_field_solver.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,16 +70,28 @@ single_field_solve!(::ClimaComms.AbstractCPUDevice, cache, x, A, b) =
_single_field_solve!(ClimaComms.device(axes(A)), cache, x, A, b)

# CPU (GPU has already called Spaces.column on arg)
_single_field_solve!(device::ClimaComms.AbstractCPUDevice, cache, x, A, b) =
Fields.bycolumn(axes(A)) do colidx
_single_field_solve_col!(
ClimaComms.device(axes(A)),
cache[colidx],
x[colidx],
A[colidx],
b[colidx],
)
function _single_field_solve!(
device::ClimaComms.AbstractCPUDevice,
cache,
x,
A,
b,
)
space = axes(x)
if space isa Spaces.FiniteDifferenceSpace
_single_field_solve_col!(device, cache, x, A, b)
else
Fields.bycolumn(space) do colidx
_single_field_solve_col!(
device,
cache[colidx],
x[colidx],
A[colidx],
b[colidx],
)
end
end
end

function _single_field_solve_col!(
::ClimaComms.AbstractCPUDevice,
Expand Down
Loading
Loading