Skip to content
This repository has been archived by the owner on Oct 31, 2023. It is now read-only.

Implement forward tensor division. #48

Open
gafter opened this issue Apr 1, 2022 · 1 comment
Open

Implement forward tensor division. #48

gafter opened this issue Apr 1, 2022 · 1 comment
Labels
good first issue Good for newcomers 🐇 performance Areas where we would like diffkt to take less time or use less memory.

Comments

@gafter
Copy link
Contributor

gafter commented Apr 1, 2022

Currently forward tensor elementwise division is implemented so that x / y is evaluated as x * y.pow(-1). That isn't very efficient, among other reasons it creates an intermediate tensor for y.pow(-1). Please implement it properly. It has already been done for scalars, and that could be used as a reference.
@gafter gafter added the 🐇 performance Areas where we would like diffkt to take less time or use less memory. label Apr 1, 2022
@gafter
Copy link
Contributor Author

gafter commented Apr 12, 2022

The same applies to reverse mode division (both for scalars and tensors)

@gafter gafter added the good first issue Good for newcomers label May 24, 2022
@gafter gafter transferred this issue from another repository May 24, 2022
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees
No one assigned
Labels
good first issue Good for newcomers 🐇 performance Areas where we would like diffkt to take less time or use less memory.
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@gafter