Back to Gap Map Infrastructure / Compilers & Model Optimization

Finch.jl

open source / Overall score: 2.5

Finch.jl is a Julia-to-Julia compiler for sparse and structured multidimensional arrays. It lets users write high-level array programs that Finch transforms behind the scenes into fused, customized sparse kernels, generalizing sparse tensor compiler techniques to a wider range of array formats.

Openness

5 high confidence

5.0

license: MIT(OSI)
source: public
core features withheld: no

MIT license body confirmed. The repository is public and unarchived and builds the whole product, and the README describes no paid tier, enterprise edition or license-gated build beside it, so source is public and the core ungated.

https://api.github.com/repos/finch-tensor/Finch.jl recorded 2026-09-02
Repo metadata for finch-jl's canonical repository - private false, archived false, fork false.
https://github.com/finch-tensor/Finch.jl/blob/main/LICENSE recorded 2026-10-07
LICENSE body is the verbatim MIT text under a Finch.jl contributors copyright line.
https://github.com/finch-tensor/Finch.jl/blob/main/README.md recorded 2026-10-07
README describes Finch as a Julia-to-Julia compiler for sparse and structured arrays that automatically generates customized, fused sparse kernels from high-level array programs, with no paid tier, enterprise edition or license-gated build beside the published source.

Adoption

1 low confidence

1.0

GitHub stars are the adoption signal for Finch.jl: the Julia General registry, where the package installs from, publishes no download counts. A separate `finch-tensor` package on PyPI is a different, Python-native library that only optionally pulls in Finch.jl as a `julia` extra, so its downloads are not counted here.

https://api.github.com/repos/finch-tensor/Finch.jl recorded 2026-09-02
Repo metadata - stargazers_count = 245 - for finch-tensor/Finch.jl.

Capability

4 medium confidence

4.0

Finch.jl is a full Julia-to-Julia compiler that lowers high-level sparse and structured array programs into fused, customized kernels. That toolchain is narrowed to one class of programs, sparse and structured tensor computation, rather than the retargetable, multi-framework, multi-vendor scope Apache TVM covers.

https://github.com/finch-tensor/Finch.jl/blob/main/README.md recorded 2026-10-07
README describes Finch as a Julia-to-Julia compiler for sparse and structured arrays that automatically generates customized, fused sparse kernels from high-level array programs, with no paid tier, enterprise edition or license-gated build beside the published source.

Verified 2026-09-27