Related articles |
---|
Frequency of integer division in source code? Peter-Lawrence.Montgomery@cwi.nl (1994-02-07) |
Re: Frequency of integer division in source code? hbaker@netcom.com (1994-02-08) |
Re: Frequency of integer division in source code? meissner@osf.org (1994-02-09) |
Re: Frequency of integer division in source code? tfj@apusapus.demon.co.uk (1994-02-10) |
Re: Frequency of integer division in source code? glew@ichips.intel.com (1994-02-11) |
Re: Frequency of integer division in source code? bazyar@netcom.com (1994-02-10) |
Re: Frequency of integer division in source code? cliffc@dawn.cs.rice.edu (1994-02-14) |
Re: Frequency of integer division in source code? chase@Think.COM (1994-02-15) |
Re: Frequency of integer division in source code? pardo@cs.washington.edu (1994-02-17) |
Newsgroups: | comp.compilers |
From: | glew@ichips.intel.com (Andy Glew) |
Keywords: | architecture, optimize |
Organization: | Intel Corp., Hillsboro, Oregon |
References: | 94-02-058 94-02-070 |
Date: | Fri, 11 Feb 1994 05:56:52 GMT |
meissner@osf.org writes:
>[re division used for constant modulus in hash table, which can be
> implemented by multiplying]
Trevor Jenkins writes:
Better yet is not to be mis-lead by Maurer's paper and to ignore his
advice by using a table whose size is a power of two. Then the integer
divide becomes an AND operation which is faster that division or
multiplication on almost all architectures.
The ``theory'' of using powers of two is explained in various papers by
Hopgood et all from the late-60/early-70s.
Powers-of-2 are, of course, the underlying principle of most hardware
cache and TLB lookups.
But they have this annoying property: seems many people like
allocating objects on, say, 64K or 1M boundaries...
--
Andy "Krazy" Glew, glew@ichips.intel.com, Intel,
M/S JF1-19, 5200 NE Elam Young Pkwy, Hillsboro, Oregon 97124-6497.
Place URGENT in email subject line for mail filter prioritization.
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.