-
Notifications
You must be signed in to change notification settings - Fork 197
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.
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
Optimize multi-threaded CKKS bootstrapping #908
Comments
I ran some experiments for CKKS bootstrapping. Results for 8 threads (on a 16-core machine):
Results for 1 core
So I only observed a speed-up of 2.1x |
In most places we rely on OMP parallelization at the DCRTPoly level. I ran a polynomial benchmark (also for N = 64K, just like for CKKS bootstrapping; see the table below) and also got a modest speed-up (8 towers vs 1): Add: 2.82x The fact that we get a low speed-up at DCRTPoly (and we have many other operations as well) leads to a small speed-up for CKKS bootstrapping.
|
Currently, we see a very modest speed-up. E.g. , when we run on a machine with 8 cores, the speed-up is only 2.2x (as compared to the single-threaded execution) vs the theoretical maximum of 8x.
The text was updated successfully, but these errors were encountered: