pashkinelfe

8 ms·

1.
by pashkinelfe 1y ago
A nice performance review for Orioledb. I'm impressed with the results. These measurements use sysbench, and complement already published benchmarks using TPC-b, and TPC-c workload: [1] https://www.orioledb.com/blog&#x2
2.
by pashkinelfe 1y ago
It's a thorough explanation for those who understand PostgreSQL internals. It probably needs some time to understand the ideas. They are elegant in my opinion. Thanks to the author!
3.
by pashkinelfe 2y ago
I'll explore performance with this parameter. Thanks for the advice!
4.
by pashkinelfe 2y ago
Author is here. Tables occupying more space than expected is a common complaint from Postgres users. I explained in simple terms what could (or should) be done to avoid this in real working conditions. And what kinds of workloads need speci
5.
Faster semantic search with HNSW indexes in pgvector(supabase.com)
7 points by pashkinelfe 3y ago|2 comments
6.
by pashkinelfe 3y ago
Pgvector 0.5.0 got HNSW indexes that is much faster for AI application than IVF in most cases. We measured pgvector performance in Supabase and present recommendations on its efficient usage. - HNSW preserves index quality even after massiv
7.
by pashkinelfe 3y ago
Interesting thing about faster distance calculations is that a very small change in the dot product calculation makes code compile to much more efficient CPU instructions and up to 2 times speedup for the whole algo. But the real jewel in t
8.
by pashkinelfe 3y ago
If you have 10M or bigger dataset of real-world OpenAI- dimensional vectors, please share, I'll use it in the next benchmarks. Random datasets are too misleading for vector search benchmarks because all ANN engines make use of internal
9.
by pashkinelfe 3y ago
Pgvector doesn't need to store everything in memory. It behaves similar to almost any Postgres AM's and store index data on disk. Performance-wise it's better to have enough memory for index data remain in shared memory buffe
10.
by pashkinelfe 3y ago
I like these measurements with ANN-benckmark! They allow to compare performance of different index implementations apples-to-apples i.e. at the same build parameters set rather than using some fixed settings (or, even worse, default setting
11.
by pashkinelfe 3y ago
I suppose that more CPU load is because more cores are loaded by doing tasks in parallel. Less CPU load is actually indicator that performance is limited by something else (IO, locks, memory coherency) that couldn't be scaled with the
12.
by pashkinelfe 4y ago
It is possible if core team decides they are better with something else. No doubt, there is some conservatism in the community.
13.
by pashkinelfe 4y ago
It's not easy to have a list of what is important to be done as there're so many different things. Generally it's good 1. To find already proposed patch of a moderate complexity and review it 2. To find a reported bug and pro
14.
by pashkinelfe 4y ago
For me it's interesting can this work without the need of external proxy server which seems to be needed only to overcome browser connections limitations and nothing else. May there be some more "internal" way to implement&#x