Article View: comp.arch
Article #12596Re: % in hash table
From: Zalman Stern
Date: Tue, 03 Oct 1995 12:06
Date: Tue, 03 Oct 1995 12:06
18 lines
1402 bytes
1402 bytes
In article <jgkDFv094.JKJ@netcom.com>, Joe Keane <jgk@netcom.com> wrote: [Using modulous to reduce a hash function to an array index is stupid.] Yep. Using a prime number modulous will mask (pun intended) a bad hash function somewhat though. > Given how slow remainder is, i like the RISC approach of calling a function. > I *like* the SPARC because the remainder function shows up in profiling, so > you know who's calling it and how much time you're wasting. Pixie on MIPS systems will pinpoint hotspot *instructions* when used properly. Hell, on a good day with a cache simulator, one could nail down cache miss hotspots using the MIPS tools. (And if you thing modulous is slow, wait 'til you see those cache misses :-)) I'm not sure whether the MIPS profiling stuff has gotten better or just rotted over the years. In the 1990 timeframe, it was state of the art. -Z-
Technical Details
Message-ID:
<zalman-0310950506260001@198.95.245.190>Path:
arc1-ca.pugleaf.net!macromedia.com!NewsWatcher!userReferences:
<43tqr6$3au@mikasa.iol.it> <449dbk$8t5@lyra.csx.cam.ac.uk> <44c0gk$jq4@blather.cs.cornell.edu> <DFLzD9.Anx@uns.bris.ac.uk> <jgkDFv094.JKJ@netcom.com>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!macromedia.com!NewsWatcher!user From: zalman@macromedia.com (Zalman Stern) Newsgroups: comp.arch,comp.arch.arithmetic,comp.lang.c Subject: Re: % in hash table Date: Tue, 03 Oct 1995 05:06:26 -0700 Organization: Macromedia Video Products Group Message-ID: <zalman-0310950506260001@198.95.245.190> References: <43tqr6$3au@mikasa.iol.it> <449dbk$8t5@lyra.csx.cam.ac.uk> <44c0gk$jq4@blather.cs.cornell.edu> <DFLzD9.Anx@uns.bris.ac.uk> <jgkDFv094.JKJ@netcom.com> NNTP-Posting-Host: 198.95.245.190