Re:Ristretto: A High Performance, Concurrent, Memory-Bound G...
by karlmcguire on Saturday September 21, 2019 @02:48PM
(#21037149)
Thank you for the kind words!
You're right about our rabbit holes. Our initial "why" for this project was essentially: "Caffeine is written in Java with excellent performance. Go should be (and usually is) competitive with Java. Why is there no Go equivalent?"
With such a high bar we were disappointed in naive, idiomatic Go solutions, but we knew there had to be some way to close the gaps.
I look forward to continuing to do so. We have a few things on the roadmap such as adaptivity (see Ben Manes' work for more info) and managing our own memory.
Re:Ristretto: A High Performance, Concurrent, Memory-Bound G...
by mrjn on Saturday September 21, 2019 @10:26PM
(#21039251)
Thanks! If not Go, we would have used C++. But the simplicity and readability of the Go codebase is a huge win. Also, C++ is just not as much fun to write in.
I joke that Go is kind of like the wild West. It comes with its Benefits and own unique challenges.
...kind of wondering, they had so at least three "well, go doesn't do ..." (lock-less data structures, thread local storage, and channels aren't high perf) rabbit holes that I wonder if go was really the right implementation language for their project (iirc they've also had articles go by about writing best-in-class go versions of rocksdb/etc.).
But, they do seem very capable of rolling their own infra, and I suppose someone has to do it to bootstrap ecosystems in each new/upcoming language.