You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The Bidding/Auction and Key/Value Servers use the same underlying library to execute User Defined Functions (i.e. code that's loaded at runtime) in JavaScript or WebAssembly. The library for this is called ROMA and it's Open Source here: https://github.com/privacysandbox/data-plane-shared-libraries
I'd like to set up some microbenchmarks that will test out how fast it is to load and execute code inside ROMA. Would anyone be comfortable having their code run in this fashion? This is very much optional!
The tradeoffs are:
Your code would have to be public and copied into the repo linked above. I'm hoping this isn't too big a deal given that code that runs on-device is already effectively public but you won't want to put anything confidential in there.
It'll get tested on a regular (multiple times a day) basis and we'll be tracking performance over time so this is a way to help make sure that library changes don't impact your code's performance.
[Editing to add: if there are particular inputs to the code that you'd like it to be benchmarked with then those would be helpful too. Hat tip @dankocoj-google.]
The text was updated successfully, but these errors were encountered:
The Bidding/Auction and Key/Value Servers use the same underlying library to execute User Defined Functions (i.e. code that's loaded at runtime) in JavaScript or WebAssembly. The library for this is called ROMA and it's Open Source here: https://github.com/privacysandbox/data-plane-shared-libraries
I'd like to set up some microbenchmarks that will test out how fast it is to load and execute code inside ROMA. Would anyone be comfortable having their code run in this fashion? This is very much optional!
The tradeoffs are:
[Editing to add: if there are particular inputs to the code that you'd like it to be benchmarked with then those would be helpful too. Hat tip @dankocoj-google.]
The text was updated successfully, but these errors were encountered: