On 11/13/22 4:23 PM, Jim Jagielski wrote:
> This article (https://vorpus.org/blog/why-does-calloc-exist/)
> looks like it's very applicable to APR, where we do the exact
> malloc-memset trick.
>
> Prelim testing on my macOS and Linux machines do show appreciable
> improvements. At the very least, maybe a compile-time flag??
>
I guess the improvements depend on the amount of memory allocated.
If you just allocate just a few pages it shouldn't matter.
Which amounts did you allocate for your tests?
Regards
Rüdiger
