Re: Table compression

henry@spsystems.net (Henry Spencer)
13 Oct 2005 20:45:19 -0400

          From comp.compilers
Related articles
[6 earlier articles]
Re: Table compression anton@mips.complang.tuwien.ac.at (2005-09-30)
Re: Table compression hannah@schlund.de (2005-09-30)
Re: Table compression cleos@nb.sympatico.ca (Cleo Saulnier) (2005-09-30)
Re: Table compression Peter_Flass@Yahoo.com (Peter Flass) (2005-10-02)
Re: Table compression paul@parsetec.com (Paul Mann) (2005-10-02)
Re: Table compression cfc@shell01.TheWorld.com (Chris F Clark) (2005-10-03)
Re: Table compression henry@spsystems.net (2005-10-13)
| List of all articles for this month |
From: henry@spsystems.net (Henry Spencer)
Newsgroups: comp.compilers
Date: 13 Oct 2005 20:45:19 -0400
Organization: SP Systems, Toronto, Canada
References: 05-09-130
Keywords: performance
Posted-Date: 13 Oct 2005 20:45:19 EDT



At the end of a message by Leonardo Teixeira Passos <leonardo@dcc.ufmg.br>
the moderator butted in:
>[Since computer memories have gotten so big, does anyone care about
>table compression any more? When your whole compiler had to fit into
>64K, compressing a few K out of the table was a big deal...


It's still an issue, but for a more subtle reason: the CPUs are getting
faster much more rapidly than the memory. Increasingly, it is worth using
a *lot* of CPU cycles to avoid a memory reference, or to keep memory
references sequential so that you make effective use of a whole cache line
before going to memory for another, or to keep tables small enough to fit
in the fastest cache.
--
spsystems.net is temporarily off the air; | Henry Spencer
mail to henry at zoo.utoronto.ca instead. | henry@spsystems.net

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.