Date: Tue, 17 Oct 1995 11:37:43 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: bde@zeta.org.au (Bruce Evans) Cc: hackers@FreeBSD.ORG, julian@ref.tfs.com Subject: Re: netisr code.. Message-ID: <199510171837.LAA28008@phaeton.artisoft.com> In-Reply-To: <199510170755.RAA00245@godzilla.zeta.org.au> from "Bruce Evans" at Oct 17, 95 05:55:47 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> That is slower at the end. However, micro-optimizations here are probably > not important. Everything except the atomic btrl could be written in C > and you probably wouldn't notice the difference. I would. ;-). I guess I'm just a computational nanosecond kinda guy. Comes from growing up on 1KHz machines with 8k or less of memory, I suppose, instead of on VAXen "where instructions are emulated and memory is free". The attitute that an optimization "doesn't matter" (this one is a bad example -- it's not really an optimization) is bad. That said, there is plenty of "low hanging fruit" that should be picked before going into instruction counting mode (unless you are doing kernel function block profiling -- if you are, then you have arrows pointing to where you need to fix the code). Now to totally confuse things: I think there should be C versions of all possible assembly code to aid porting. Anything that works is better than anything that doesn't. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199510171837.LAA28008>