AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Importance of Parallelism in Parallel Processing
There was big improvement to reduce contention how we've worked with the buffer pool is organized. If you have more sessions, active sessions, than cores, it's not a good position already. I remember a great work from in post, blog post from Alexander Korotkoff towards 1 million TPS on single node. It was interesting that it was done around the same time when MySQL did it as well. But it's not related to processing of single queries related to multi-user, multi-session processing.