Back to Gap Map Infrastructure / Compilers & Model Optimization

TACO

open source / Overall score: 3.0

TACO (the Tensor Algebra Compiler) is a C++ library and code generator, developed at MIT CSAIL and Adobe, that compiles sparse and dense tensor algebra expressions into native code competitive with hand-optimized kernels, using novel sparse iteration-space compiler techniques.

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/tensor-compiler/taco recorded 2026-09-02
Repo metadata for taco's canonical repository - private false, archived false, fork false.
https://github.com/tensor-compiler/taco/blob/master/LICENSE recorded 2026-10-07
LICENSE body carries the verbatim MIT text under a copyright line naming MIT CSAIL, Adobe Systems and other contributors.
https://github.com/tensor-compiler/taco/blob/master/README.md recorded 2026-10-07
README describes taco as a C++ library that computes sparse and dense tensor algebra expressions with performance competitive with hand-optimized kernels, usable as a library or as a C code generator, with no paid tier, enterprise edition or license-gated build beside the published source.

Adoption

2 low confidence

2.0

The `taco` package on PyPI belongs to an unrelated project by a different author, and TACO itself ships only as a source build, so GitHub stars are the only adoption signal available.

https://api.github.com/repos/tensor-compiler/taco recorded 2026-09-02
Repo metadata - stargazers_count = 1,367 - for tensor-compiler/taco.

Capability

4 medium confidence

4.0

TACO is a full compilation toolchain - parser through code generation - for sparse and dense tensor algebra expressions, targeting both CPUs and GPUs. That toolchain is narrowed to one class of programs rather than the retargetable, multi-framework, multi-vendor scope Apache TVM covers.

https://github.com/tensor-compiler/taco/blob/master/README.md recorded 2026-10-07
README describes taco as a C++ library that computes sparse and dense tensor algebra expressions with performance competitive with hand-optimized kernels, usable as a library or as a C code generator, with no paid tier, enterprise edition or license-gated build beside the published source.

Verified 2026-09-27