On Sat, Mar 27, 2021 at 4:52 AM Andrey Borodin <x4mmm@yandex-team.ru> wrote:
> Some thoughts on HashTable patch:
> 1. Can we allocate bigger hashtable to reduce probability of collisions?
Yeah, good idea, might require some study.
> 2. Can we use specialised hashtable for this case? I'm afraid hash_search() does comparable number of CPU cycles as
simplecycle from 0 to 128. We could inline everything and avoid hashp->hash(keyPtr, hashp->keysize) call. I'm not
insistingon special hash though, just an idea.
I tried really hard to not fall into this rabbit h.... [hack hack
hack], OK, here's a first attempt to use simplehash, Andres's
steampunk macro-based robinhood template that we're already using for
several other things, and murmurhash which is inlineable and
branch-free. I had to tweak it to support "in-place" creation and
fixed size (in other words, no allocators, for use in shared memory).
Then I was annoyed that I had to add a "status" member to our struct,
so I tried to fix that. Definitely needs more work to think about
failure modes when running out of memory, how much spare space you
need, etc.
I have not experimented with this much beyond hacking until the tests
pass, but it *should* be more efficient...
> 3. pageno in SlruMappingTableEntry seems to be unused.
It's the key (dynahash uses the first N bytes of your struct as the
key, but in this new simplehash version it's more explicit).