Go: What We Got Right, What We Got Wrong

Re:Go: What We Got Right, What We Got Wrong by pcwalton on Thursday January 04, 2024 @06:02PM (#38873526)

I usually just say that Go implements "userspace threading", since that's really what it is. Some early, pre-pthreads, implementations of Linux threads worked the same way as Go does and they usually called such implementations "M:N", to indicate M userspace threads mapped onto N kernel threads, so "M:N" is a good descriptor too.

[ Parent | on HN ]

Re:Go: What We Got Right, What We Got Wrong by dekhn on Thursday January 04, 2024 @06:30PM (#38873719)

but that's not really accurate. It uses system threads. Userspace threading doesn't make a transition to kernel (IIUC).

[ Parent | on HN ]

Re:Go: What We Got Right, What We Got Wrong by pcwalton on Thursday January 04, 2024 @06:41PM (#38873784)

Go doesn't need to transition to the kernel to switch threads, does it? It just saves and loads the register state in userland.

[ Parent | on HN ]

Re:Go: What We Got Right, What We Got Wrong by dekhn on Thursday January 04, 2024 @06:51PM (#38873871)

IIUC it can, sometimes, avoid a kernel transition when it knows it can schedule the recipient of a message, but I believe that golang creates a threadpool for running goroutines on platforms that use thread primitives.

From https://github.com/golang/go/blob/c7d6c6000a84b61ac8bb2e38e8... I believe "worker threads" are OS threads.

[ Parent | on HN ]