Xref: utzoo comp.lang.c:14491 comp.lang.misc:2269 Path: utzoo!utgpu!watmath!clyde!att!rutgers!gatech!ukma!nrl-cmf!ames!haven!ncifcrf!nlm-mcs!adm!smoke!gwyn From: gwyn@smoke.BRL.MIL (Doug Gwyn ) Newsgroups: comp.lang.c,comp.lang.misc Subject: Re: instruction timings (was: Assembly or ....) Message-ID: <9045@smoke.BRL.MIL> Date: 2 Dec 88 17:17:00 GMT References: <1388@aucs.UUCP> <729@convex.UUCP> <1961@crete.cs.glasgow.ac.uk> <1988Nov29.181235.23628@utzoo.uucp> <960@vsi.COM> Reply-To: gwyn@brl.arpa (Doug Gwyn (VLD/VMB)) Organization: Ballistic Research Lab (BRL), APG, MD. Lines: 12 In article <960@vsi.COM> friedl@vsi.COM (Stephen J. Friedl) writes: >Why is this? When I was hacking on the VAX, nobody could ever >tell me how long anything took, and empirical measurements were >pretty tedious. Is it laziness on the vendor's part or are there >good reasons for this? Exact computations of instruction timings, already quite difficult for the VAX-11/780, have become so context-dependent as to be not worth attempting. Time taken for a given instruction depends on such things as cache hit rate, fault rate, instruction mix, and so forth. What really matters is overall system performance, which is a statistical matter.