Twitter | Pretraživanje | |
Glauber Costa
Wondering how are the early results for the io_uring backend for seastar? 50% faster in the first benchmark (workload is small 512-byte reads with iodepth of one, competing for dispatch time against a CPU-bound constant workload)
Reply Retweet Označi sa "sviđa mi se" More
Mark Callaghan 29. sij
Odgovor korisniku/ci @glcst
Do you have workloads that do 512-byte reads? If so that is great -- do more IO with less CPU.
Reply Retweet Označi sa "sviđa mi se"
Glauber Costa 29. sij
Odgovor korisniku/ci @MarkCallaghanDB
yes, although it is rare for Scylla. For seastar, the framework, though, people use it for god knows what so I like to test the extremes
Reply Retweet Označi sa "sviđa mi se"