Conversation
meta-cla
bot
added
the
CLA Signed
Contributor
facebook-github-bot
pushed a commit
that referenced
this pull request
Feb 2, 2026
Summary: Delete SubsetGaussNewtonSolver as all of its functionality has been absorbed by the standard GaussNewtonSolver. The subset optimization capability that SubsetGaussNewtonSolver provided is now natively supported by GaussNewtonSolver, making this specialized solver redundant. This removes: - SubsetGaussNewtonSolver class and implementation - Associated unit tests - Python bindings for SubsetGaussNewtonSolver Reviewed By: cstollmeta Differential Revision: D91914618
facebook-github-bot
force-pushed
the
export-D91914618
branch
from
e4cd955 to
90bf7d4
Compare
facebook-github-bot
pushed a commit
that referenced
this pull request
Feb 3, 2026
Summary: Delete SubsetGaussNewtonSolver as all of its functionality has been absorbed by the standard GaussNewtonSolver. The subset optimization capability that SubsetGaussNewtonSolver provided is now natively supported by GaussNewtonSolver, making this specialized solver redundant. This removes: - SubsetGaussNewtonSolver class and implementation - Associated unit tests - Python bindings for SubsetGaussNewtonSolver Reviewed By: cstollmeta Differential Revision: D91914618
facebook-github-bot
force-pushed
the
export-D91914618
branch
from
90bf7d4 to
c4016b1
Compare
facebook-github-bot
pushed a commit
that referenced
this pull request
Feb 3, 2026
Summary: Delete SubsetGaussNewtonSolver as all of its functionality has been absorbed by the standard GaussNewtonSolver. The subset optimization capability that SubsetGaussNewtonSolver provided is now natively supported by GaussNewtonSolver, making this specialized solver redundant. This removes: - SubsetGaussNewtonSolver class and implementation - Associated unit tests - Python bindings for SubsetGaussNewtonSolver Reviewed By: cstollmeta, jeongseok-meta Differential Revision: D91914618
facebook-github-bot
force-pushed
the
export-D91914618
branch
from
c4016b1 to
6d0bbc5
Compare
facebook-github-bot
pushed a commit
that referenced
this pull request
Feb 3, 2026
Summary: Delete SubsetGaussNewtonSolver as all of its functionality has been absorbed by the standard GaussNewtonSolver. The subset optimization capability that SubsetGaussNewtonSolver provided is now natively supported by GaussNewtonSolver, making this specialized solver redundant. This removes: - SubsetGaussNewtonSolver class and implementation - Associated unit tests - Python bindings for SubsetGaussNewtonSolver Reviewed By: cstollmeta, jeongseok-meta Differential Revision: D91914618
facebook-github-bot
force-pushed
the
export-D91914618
branch
from
6d0bbc5 to
add1f8f
Compare
facebook-github-bot
pushed a commit
that referenced
this pull request
Feb 4, 2026
Summary: Delete SubsetGaussNewtonSolver as all of its functionality has been absorbed by the standard GaussNewtonSolver. The subset optimization capability that SubsetGaussNewtonSolver provided is now natively supported by GaussNewtonSolver, making this specialized solver redundant. This removes: - SubsetGaussNewtonSolver class and implementation - Associated unit tests - Python bindings for SubsetGaussNewtonSolver Reviewed By: cstollmeta, jeongseok-meta Differential Revision: D91914618
facebook-github-bot
force-pushed
the
export-D91914618
branch
from
add1f8f to
3be1225
Compare
Summary: Update ResizeableMatrix to use AlignedAllocator with 128-byte alignment instead of std::vector's default alignment. This improves cache line efficiency and SIMD performance by ensuring data is properly aligned for AVX-512 operations and avoiding false sharing in multi-threaded scenarios. Reviewed By: yutingye Differential Revision: D92361055
…version 2) (#1001) Summary: Computing JtJ is completely bottlenecked by the actual matrix multiplication, and this is highly dependent on maximizing the throughput of L3 BLAS operations. Currently we are filling in the J matrix a block at a time and then computing JtJ on that block, but by using the full J matrix we leave large gaps in the effective J matrix that limit how much of it can fit into cache. If we instead pack everything tightly into a smaller matrix we may be able to exploit better cache locality. The reason we use ResizeableMatrix for this is that the regular Eigen matrix class _always_ reallocates every time you resize, which would eliminate all possible speedups. Reviewed By: yutingye Differential Revision: D92343862
Summary: Moved the `padJacobianRows` function to a centralized location in `solver_function.h` as `padToSimdAlignment`. This eliminates code duplication across multiple solver files that each had their own local copy of the same padding logic. The function pads row counts to multiples of 8 for SIMD alignment, which is required for AVX operations on Jacobian matrices. Reviewed By: jeongseok-meta Differential Revision: D92426600
Summary: Extracted ResizeableMatrix from online_householder_qr.h into its own header file (resizeable_matrix.h) to reduce dependencies. This allows solver_function.h to include only the lightweight ResizeableMatrix class without pulling in all the QR solver machinery. ResizeableMatrix is a utility class that provides matrix storage that can be resized without reallocating when shrinking, which is useful for block-wise Jacobian computations where the block size varies. Reviewed By: jeongseok-meta Differential Revision: D92427665
Summary: Delete SubsetGaussNewtonSolver as all of its functionality has been absorbed by the standard GaussNewtonSolver. The subset optimization capability that SubsetGaussNewtonSolver provided is now natively supported by GaussNewtonSolver, making this specialized solver redundant. This removes: - SubsetGaussNewtonSolver class and implementation - Associated unit tests - Python bindings for SubsetGaussNewtonSolver Differential Revision: D91914618
facebook-github-bot
force-pushed
the
export-D91914618
branch
from
3be1225 to
3910098
Compare
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
Summary:
Delete SubsetGaussNewtonSolver as all of its functionality has been absorbed by the
standard GaussNewtonSolver. The subset optimization capability that SubsetGaussNewtonSolver
provided is now natively supported by GaussNewtonSolver, making this specialized solver
redundant.
This removes:
Differential Revision: D91914618