Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: size of a datatype

  • Subject: Re: size of a datatype
  • From: Bob Hauck <postmaster@xxxxxxxxxxxxxxxxxxxxx>
  • Date: Sun, 14 May 2006 20:43:27 -0400
  • Newsgroups: comp.os.linux.advocacy
  • Organization: Hauck Family
  • References: <1147499965.060218.303410@i40g2000cwc.googlegroups.com> <40egj3-uju.ln1@ridcully.fsnet.co.uk> <1147506449.314246.192970@j73g2000cwa.googlegroups.com> <bto9g.1919$x4.378@newsread3.news.pas.earthlink.net> <Srq9g.68749$wl.13434@text.news.blueyonder.co.uk> <4682456.jc0GDkRpje@schestowitz.com>
  • Reply-to: bobh at the domain haucks dot org
  • User-agent: slrn/0.9.8.1 (Debian)
  • Xref: news.mcc.ac.uk comp.os.linux.advocacy:1109207
On Sun, 14 May 2006 04:16:28 +0100, Roy Schestowitz
<newsgroups@xxxxxxxxxxxxxxx> wrote:

>                 Source: http://en.wikipedia.org/wiki/128-bit
>
> Can't wait 'till the kilobit processor. Imagine the compexity of the
> chip and the compiler...

There are some DSP's that use what's called a "Very Long Instruction
Word" architecture, where multiple instructions are packed into a 64-bit
(or longer) word.  The Crusoe processor from Transmeta also used this
architecture, for a fairly well-known example.

VLIW basiclly moves responsibility for the discovery of parallel
instructions from the hardware (as in a superscaler design) to the
compiler.  This actually makes the VLIW simpler than it would otherwise
be but makes the compiler more complicated.


-- 
 -| Bob Hauck
 -| A proud member of the unhinged moonbat horde.
 -| http://www.haucks.org/

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index