Re: complex instruction set compiler

Tom Moog <tmoog@mcs.net>
20 Mar 1998 01:18:45 -0500

          From comp.compilers
Related articles
complex instruction set compiler vjain@ececs.uc.edu (Vipin Jain) (1998-03-15)
Re: complex instruction set compiler ian@five-d.com (1998-03-15)
Re: complex instruction set compiler chase@naturalbridge.com (David Chase) (1998-03-15)
Re: complex instruction set compiler Eric.Hildum@Japan.NCR.COM (Eric Hildum) (1998-03-18)
Re: complex instruction set compiler richard@cogsci.ed.ac.uk (1998-03-18)
Re: complex instruction set compiler tmoog@mcs.net (Tom Moog) (1998-03-20)
Re: complex instruction set compiler jan@mailhost.neuroinformatik.ruhr-uni-bochum.de (Jan Vorbrueggen) (1998-03-20)
| List of all articles for this month |
From: Tom Moog <tmoog@mcs.net>
Newsgroups: comp.compilers,comp.arch
Date: 20 Mar 1998 01:18:45 -0500
Organization: Polhode Inc
References: 98-03-144 98-03-149 98-03-175
Keywords: architecture, performance

The POLY instruction was actually slower than the equivalent simple
operations on some models of VAX. The first model, the 780 had an
INDEX instruction to do most of the work of array indexing and bounds
checking. It was actually slower than the equivalent set of simpler
instructions because the hardware floating point accelerator couldn't
be used to perform the multiply step (the FPA performed integer
multiply and divide). In recent years the POLY instructions were
actually removed from many of the math library functions, probably
because the newer machines emulated POLY, making it super slow.


Tom Moog
--

Post a followup to this message

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