In lookup and appendInternal, it's not actually necessary to check the validity mask for anything other than determining the number of valid entries.
I think this makes a slight difference, but the variance between test runs is large enough in comparison to any performance improvement that it's very hard to tell.
In lookup and appendInternal, it's not actually necessary to check the validity mask for anything other than determining the number of valid entries. I think this makes a slight difference, but the variance between test runs is large enough in comparison to any performance improvement that it's very hard to tell.