Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Vectorized RNG #405

Open
mratsim opened this issue Nov 17, 2019 · 0 comments
Open

Vectorized RNG #405

mratsim opened this issue Nov 17, 2019 · 0 comments

Comments

@mratsim
Copy link
Owner

mratsim commented Nov 17, 2019 •
edited
Loading

randomTensor uses the default rng from Nim stdlib which is xoroshiro128.

Through SIMD vectorization, it's possible to improve xoroshiro256 by 4X see JuliaLang/julia#27614 which would be quite useful when initializing tensors of millions of elements.

See also http://prng.di.unimi.it/xoshiro256+-vect-speed.c and http://prng.di.unimi.it/

Note that the xoshiro family provides "jump sequence for parallel computation http://prng.di.unimi.it/xoshiro256plus.c
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@mratsim