export SUnitRange #474
export SUnitRange #474
Conversation
|
Yes, this isn't unreasonable... :) There has been a long-standing plan to support static ranges, LinearIndices and CartesianIndices with StaticArrays. These things would be returned by |
|
Sure, if it's a question of API stability/not committing to this exact form for fewer "breaking" changes once more general ranges are worked out I'm happy to keep manually importing this secret functionality. |
|
OK thanks. Please bump this again soon-ish if I don't get to this. |
c42f
added
feature
|
This PR addresses #620. It would be nice to have a docstring though, perhaps with an example in use, and the docs should also mention this. |
|
Friendly bump --- could this be merged please? |
1537: Use static range in Grad setproperty! r=mwarusz a=mwarusz # Description After #1458 I noticed a slowdown. This PR proposes a fix. Running ```julia --project=. experiments/AtmosLES/bomex_les.jl --monitor-timestep-duration 100steps``` I get Before: ``` ┌ Info: Wall-clock time per time-step (statistics across MPI ranks) │ maximum (s) = 1.5273617989999998e-02 │ minimum (s) = 1.5273617989999998e-02 │ median (s) = 1.5273617989999998e-02 └ std (s) = NaN ``` After: ``` Info: Wall-clock time per time-step (statistics across MPI ranks) │ maximum (s) = 1.1554147779999999e-02 │ minimum (s) = 1.1554147779999999e-02 │ median (s) = 1.1554147779999999e-02 └ std (s) = NaN ``` The fix relies on unexported `StaticArrays` API but JuliaArrays/StaticArrays.jl#474 makes me think that this is ok. 1538: Rename step (intrinsic func) r=charleskawczynski a=charleskawczynski # Description Closes #95. Co-authored-by: Maciej Waruszewski <[email protected]> Co-authored-by: Charles Kawczynski <[email protected]>
1537: Use static range in Grad setproperty! r=mwarusz a=mwarusz # Description After #1458 I noticed a slowdown. This PR proposes a fix. Running ```julia --project=. experiments/AtmosLES/bomex_les.jl --monitor-timestep-duration 100steps``` I get Before: ``` ┌ Info: Wall-clock time per time-step (statistics across MPI ranks) │ maximum (s) = 1.5273617989999998e-02 │ minimum (s) = 1.5273617989999998e-02 │ median (s) = 1.5273617989999998e-02 └ std (s) = NaN ``` After: ``` Info: Wall-clock time per time-step (statistics across MPI ranks) │ maximum (s) = 1.1554147779999999e-02 │ minimum (s) = 1.1554147779999999e-02 │ median (s) = 1.1554147779999999e-02 └ std (s) = NaN ``` The fix relies on unexported `StaticArrays` API but JuliaArrays/StaticArrays.jl#474 makes me think that this is ok. Co-authored-by: Maciej Waruszewski <[email protected]>
|
What's the status of this PR? Is this just waiting for resolving the conflicts? |
|
Friendly ping; please reconsider this PR. If anything else needs to be done I am willing to do it or make another PR if that is preferred. |
|
I think the only problem with Considering |
I feel like
SUnitRangeis useful enough (especially for slicing otherStaticArrays) to merit exporting.