pymc-devs / pymc-devs/pytensor
Group sum of vector inner products with same vector
Open
Nobody has claimed this yet.
graph rewriting
linalg
performance
- Dominant language
- Python
- Stars
- 644
- Forks
- 208
- Avg merge
- 2d 14h
- Merged PRs (30d)
- 16
Description
Description
import numpy as np
x1, x2, y = np.random.normal(size=(3, 1000, 1))
assert np.isclose(x1.T @ y + x2.T @ y, (x1 + x2).T @ y)
%timeit x1.T @ y + x2.T @ y
%timeit (x1 + x2).T @ y
1.86 μs ± 21.1 ns per loop (mean ± std. dev. of 7 runs, 1,000,000 loops each)
1.33 μs ± 13.9 ns per loop (mean ± std. dev. of 7 runs, 1,000,000 loops each)
Works with either shared y, or shared x.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the Python/NumPy reproducer in the issue and inspect PyTensor's expression-optimization passes for handling repeated inner products. Confirm the transformation preserves the shown equality for shared x or y, then compare the grouped expression's performance with the two separate products.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- numpy, python
- Domain
- compilers, performance
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100