Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!mcsun!ukc!dcl-cs!aber-cs!pcg
From: pcg@aber-cs.UUCP (Piercarlo Grandi)
Newsgroups: comp.arch
Subject: Re: Number of registers, MIPS, GNU cc
Message-ID: <2200@aber-cs.UUCP>
Date: 9 Jan 91 19:29:35 GMT
Reply-To: pcg@cs.aber.ac.uk (Piercarlo Grandi)
Organization: Dept of CS, UCW Aberystwyth
	(Disclaimer: my statements are purely personal)
Lines: 36


    I (pcg) had written:

  	name	total	spare	preserved
  
  	U12	12*	5	5
  	U16	16*	9	9
  	U24	24*	17	9
  	U32	32	25	9

In article <37605@cup.portal.com> ts@cup.portal.com (Tim W Smith) writes:
  
  How come spare + preserved - total is +-2 all the time rather
  than 0?  What are the other -+2 registers?

Sorry for not being clear here: 'preserved' and 'spare' overlap. For example
U24 has 32 real registers, 8 are marked unavailable, 7 are marked fixed use
(really the same as the previous 8), 17 are marked available, and 9 of these
17 are preserved across function calls (callee saves). Have a look at the
tail of the original message, where the definitions of the relevant arrays
are, and everything will be clear.

I have done some further experiments that show that the number of registers
saved by the callee or by the caller does not matter, at least in this case.

Incidentally: I have measured, for fun, also the time taken by the GNU cpp
on the same source. There number of registers seems to make a far more
pronounced difference than for GNU cc1. This is highly perplexing, so I am
looking into it -- will report, eventually.

Finally: if comebody has some substantial floating point program(s), it
would be interesting to see hwo number of registers affect them.
-- 
Piercarlo Grandi                   | ARPA: pcg%uk.ac.aber.cs@nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth        | UUCP: ...!mcsun!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk