Skip to content

Conversation

@devnexen
Copy link
Member

No description provided.

some minor changes but above all trying to address the little concerns
regarding errors bookkeeping.
@ndossche
Copy link
Member

It's often a good idea to choose a factor of 1.5, there's a mathematical reason behind it (https://github.com/facebook/folly/blob/main/folly/docs/FBVector.md#memory-handling).

@devnexen
Copy link
Member Author

yes will try ; even though usage and growth might not be that as critical as your (nice) link contexts.

@Girgias
Copy link
Member

Girgias commented Nov 23, 2025

It's often a good idea to choose a factor of 1.5, there's a mathematical reason behind it (https://github.com/facebook/folly/blob/main/folly/docs/FBVector.md#memory-handling).

Interesting, we currently extend arrays by a factor of 2, I wonder what would happen if we do 1.5 instead.

@ndossche
Copy link
Member

Interesting, we currently extend arrays by a factor of 2, I wonder what would happen if we do 1.5 instead.

Won't work because the hash masking depends on it being a power of 2.

@devnexen devnexen marked this pull request as ready for review November 23, 2025 13:55
@devnexen devnexen requested a review from dstogov as a code owner November 23, 2025 13:55
using just shift with integer rather than scalar
integer to float conversion (shlr vs cvtsi2sd then mulsd on intel)
@TimWolla
Copy link
Member

Can we get a clearer PR title?

@devnexen devnexen changed the title Zend little upd Nov 24, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

4 participants