[Proof of Concept] Implement data type to define threaded broadcasting #2284
+75
−4
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
efaulhaber
added
enhancement
Review checklistThis checklist is meant to assist creators of PRs (to let them know what reviewers will typically look for) and reviewers (to guide them in a structured review process). Items do not need to be checked explicitly for a PR to be eligible for merging. Purpose and scope
Code quality
Documentation
Testing
Performance
Verification
Created with ❤️ by the Trixi.jl community. |
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #2284 +/- ##
===========================================
- Coverage 89.49% 79.09% -10.40%
===========================================
Files 490 490
Lines 39507 39536 +29
===========================================
- Hits 35355 31270 -4085
- Misses 4152 8266 +4114
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
ranocha
reviewed
Feb 14, 2025
| @@ -101,8 +102,78 @@ function semidiscretize(semi::AbstractSemidiscretization, tspan; | |||
| return ODEProblem{iip, specialize}(rhs!, u0_ode, tspan, semi) | |||
| end | |||
|
|
|||
| struct ThreadedBroadcastArray{T, N, A} <: AbstractArray{T, N} | |||
There was a problem hiding this comment.
Suggested change
| struct ThreadedBroadcastArray{T, N, A} <: AbstractArray{T, N} | |
| struct ThreadedBroadcastArray{T, N, A <: AbstractArray{T, N}} <: AbstractArray{T, N} |
Comment on lines
+117
to
+118
| Base.parent(m::ThreadedBroadcastArray) = m.array | ||
| Base.size(m::ThreadedBroadcastArray) = size(m.array) |
There was a problem hiding this comment.
Suggested change
| Base.parent(m::ThreadedBroadcastArray) = m.array | |
| Base.size(m::ThreadedBroadcastArray) = size(m.array) | |
| Base.parent(m::ThreadedBroadcastArray) = m.array | |
| Base.pointer(m::ThreadedBroadcastArray) = pointer(parent(m)) | |
| Base.size(m::ThreadedBroadcastArray) = size(parent(m)) |
It's also on the agenda for Tuesday. |
|
Also see trixi-framework/TrixiParticles.jl#722 for the TrixiParticles version of this. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
As discussed in #2283.
The new data type
ThreadedBroadcastArraywraps any other array, but redefines broadcasting and functions likecopyto!andfill!to use@threadedfrom Trixi.jl.Using this data type for
u_odemakes broadcasting inperform_step!of the time integration multithreaded.This has the same effect as setting
thread = True()for time integration schemes that support this option, but it also works for other schemes and would follow a potential redefinition of@threadedaway from Polyester.jl.Without threaded time integration:
With
thread = True():With this PR:
Some notes:
thread = True()and this PR is within measuring errors.perform_step!of the Carpenter-Kennedy scheme.@batchapproach always allocates a little. Interestingly, thethread = True()version, which uses FastBroadcast.jl (which in turn is based on Polyester.jl), does not allocate for thebroadcastXtimers. Not sure why. This is the code they're using:https://github.com/YingboMa/FastBroadcast.jl/blob/ad586d83ffcac15c92969b93dd5cf0c8fd025af9/src/FastBroadcast.jl#L326-L334