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.

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

Start vectorizing TensorPrimitives #91596

Merged
merged 5 commits into from
Sep 7, 2023

Conversation

stephentoub
Copy link
Member

@stephentoub stephentoub commented Sep 5, 2023

Note I added an ifdef for NET8_0_OR_GREATER because my assumption is all of this is going to be backported to the release/8.0 branch, at which point we'll also be building for net7.0.

Just does two functions to establish the files into which the rest of the implementations can be moved.
@ghost
Copy link

ghost commented Sep 5, 2023

Tagging subscribers to this area: @dotnet/area-system-numerics
See info in area-owners.md if you want to be subscribed.

Issue Details

Just does two functions to establish the files into which the rest of the implementations can be moved, so we can work on this in parallel on the rest.

If you had something else in mind, just let me know.

(Note I added an ifdef for NET8_0_OR_GREATER because my assumption is all of this is going to be backported to the release/8.0 branch, at which point we'll also be building for net7.0.)

Author: stephentoub
Assignees: -
Labels:

area-System.Numerics

Milestone: -

@stephentoub
Copy link
Member Author

(I'm revamping as well, to share more code between the various operations. Will post an update shortly...)

@stephentoub
Copy link
Member Author

stephentoub commented Sep 5, 2023

I pushed a new commit that:

  • Handles most of the other operations that were there. It currently leaves out the log/exp/trig functions.
  • Shares the implementations using an interface and either static abstract interface methods (in .NET Core) or an instance-based interface with a dummy parameter passed (in .NET Standard)
  • Updates the logic to handle the trailing elements with a single final vector rather than cascading through smaller-width vectors

I did not add any logic for alignment. I'll leave that for you, Tanner, for a subsequent change if you think it's valuable enough.

- Share public methods between implementations
- Reduce boilerplate code in netstandard impl with AsVector helper
- Align arguments
- Add some missing tests based on code coverage
- Add argument name to throw helper
@stephentoub
Copy link
Member Author

failure is #76454

@stephentoub stephentoub merged commit 07000a7 into dotnet:main Sep 7, 2023
103 of 105 checks passed
@stephentoub stephentoub deleted the vectorizetensor branch September 7, 2023 10:36
michaelgsharp pushed a commit to michaelgsharp/runtime that referenced this pull request Sep 18, 2023
* Start vectorizing TensorPrimitives

Just does two functions to establish the files into which the rest of the implementations can be moved.
ericstj pushed a commit that referenced this pull request Sep 19, 2023
* added Bcl.Numerics

* Adding a naive implementation of various primitive tensor operations (#91228)

* Adding a naive implementation of various primitive tensor operations

* Adding tests covering the new tensor primitives APIs

* Adding tensor primitives APIs to the ref assembly

* Allow .NET Framework to build/run

* Sync TFMs between ref and src, csproj simplication and clean-up

* Apply suggestions from code review

Co-authored-by: Viktor Hofer <[email protected]>

* Don't use var

* Fix the S.N.Tensors readme and remove the file marking it as non-shipping

---------

Co-authored-by: Viktor Hofer <[email protected]>
Co-authored-by: Michael Sharp <[email protected]>

* Start vectorizing TensorPrimitives (#91596)

* Start vectorizing TensorPrimitives

Just does two functions to establish the files into which the rest of the implementations can be moved.

* 6 more naive methods for Tensor Primitives. (#92142)

* 6 more naive methods

* updates from pr comments

* Add remaining set of TensorPrimitives APIs for .NET 8 (#92154)

* Add remaining set of TensorPrimitives APIs for .NET 8

Adds non-vectorized implementations of:
- Max
- Min
- MaxMagnitude
- MinMagnitude
- IndexOfMax
- IndexOfMin
- IndexOfMaxMagnitude
- ConvertToHalf (only on .NET Core)
- ConvertToSingle (only on .NET Core)
- IndexOfMinMagnitude

Adds vectorized implementations of:
- Sum
- SumOfSquares
- SumOfMagnitudes
- Product
- ProductOfSums
- ProductOfDifferences

Also includes the helpers that'll make it trivial to vectorize Dot.

Beyond vectorizing the non-vectorized ones, the vectorized implementations should be improved further, including:
- Handling alignment better
- Vectorizing the remainder that doesn't fit in a vector rather than falling back to scalar

* Cleanup after previous PR, vectorize CosineSimilarity/Dot/L2Normalize/Distance, add tests

* Address PR feedback, and fix a few other issues

* Fix TensorPrimitives.CosineSimilarity to use vectorized implementations (#92204)

* Fixed duplicated code from merge.

* New Microsoft.BCL.Numerics package (#91074)

* bcl numberics library added

* bcl done

* added explicit 2.1 target

* Minor doc updates

* Apply suggestions from code review

Co-authored-by: Viktor Hofer <[email protected]>

* fixes from PR comments

* minor csproj fixes

* fixed ref target frameworks

* minor ref csproj updates

* minor csproj updates

---------

Co-authored-by: Viktor Hofer <[email protected]>

* Microsoft.Bcl.Numerics.Tests: fix restore failure when DotNetBuildFromSource. (#91402)

* Microsoft.Bcl.Numerics.Tests: fix restore failure when DotNetBuildFromSource.

* Use NetCoreAppCurrent.

* Try fix CI test failures.

---------

Co-authored-by: Tanner Gooding <[email protected]>
Co-authored-by: Viktor Hofer <[email protected]>
Co-authored-by: Stephen Toub <[email protected]>
Co-authored-by: Tom Deseyn <[email protected]>
@ghost ghost locked as resolved and limited conversation to collaborators Oct 7, 2023
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants