28.10.2026 08:12, Kirill Reshke wrote: > On Mon, 26 Oct 2026 at 11:42, Roman Khapov <[email protected]> wrote: >> >> Hi hackers! >> >> While working on [1] I run performance tests on different environments, >> including laptop with macOS. pgbench is usually chosen for this, >> but on one of the last runs I discovered that it uses select() on macOS >> (which lacks >> the ppoll() syscall), which caps the max amount of clients at 1024 >> (FD_SETSIZE). > > Hi! pgbench with this many clients is an uncommon case, but I can see > the benefit in stress-testing connection allocation, which in case of > connection pooler is of interest. So overall +1 for supporting this
We regularly test with couple of thousand clients. But we have to set -j matching number of -c because otherwise pgbench is bottleneck itself. Partially it is because of using ppoll, partially because some algorithms are O(n) on every wakeup in pgbench. pgbench definitely needs more love. -- regards Yura Sokolov aka funny-falcon
