I am working on ultra-low latency emulation. I have a RISC-V emulator (https://github.com/fwsGonzo/libriscv) that I am creating to fork itself really fast and use almost no working memory. Programs can make do with 64kb memory, which includes the emulator itself and everything it uses.
Those measurements are from a production environment, meaning these numbers are very real! In a synthetic benchmark the fork happens at just ~200 nanoseconds, and it's really a meaningless number.
https://cloud.nwcs.no/index.php/s/iP6aYJaoBtXbbqM
Those measurements are from a production environment, meaning these numbers are very real! In a synthetic benchmark the fork happens at just ~200 nanoseconds, and it's really a meaningless number.