home *** CD-ROM | disk | FTP | other *** search
- Path: mail2news.demon.co.uk!genesis.demon.co.uk
- From: Lawrence Kirby <fred@genesis.demon.co.uk>
- Newsgroups: comp.lang.c
- Subject: Re: Indexed addressing mode on MIPS R8000
- Date: Sat, 20 Jan 96 01:21:22 GMT
- Organization: none
- Message-ID: <822100882snz@genesis.demon.co.uk>
- References: <4dohb1$ccr@elna.ethz.ch>
- Reply-To: fred@genesis.demon.co.uk
- X-NNTP-Posting-Host: genesis.demon.co.uk
- X-Newsreader: Demon Internet Simple News v1.27
- X-Mail2News-Path: genesis.demon.co.uk
-
- In article <4dohb1$ccr@elna.ethz.ch> chagmann "Christian Hagmann" writes:
-
- >I use some Base-pointers and some Offset-pointers in C. I think, the Compiler
- >should translate an expression like 'a0=*(b0+off0);' as an indexed addressing
- >mode 'LDXC1 $fx,$y,$z'. Sometimes it is translated in this way, but sometimes
- >it is translated in the way, that b0 and off0 become added (Integer-Operation)
- >and loaded in the Offset addressing mode 'LDC1 $fx,0($q)' with $q=$y+$z. This
- >disturbs the flow of the programm and makes it slow.
-
- Have you benchmarked it to determine that it is actually slower? Anyway
- so long as the output is correct the requirements of the C language are
- satisfied. Performance is purely a quality of implementation issue. If you
- have a problem with the code quality of a particular compiler you should
- take it up with the manufacturer. The place to discuss it on Usenet is
- a newsgroup related to the platform in question, maybe in the comp.sys.sgi
- hierarchy in this case.
-
- >So, my question is:
- >
- >Are there any restrictions in register-register indexed addressing Mode on the
- >MIPS R8000 or does the IRIX restrict this addressing Mode?
-
- Your question is extremely system specific and really has nothing to do
- with the C language.
-
- --
- -----------------------------------------
- Lawrence Kirby | fred@genesis.demon.co.uk
- Wilts, England | 70734.126@compuserve.com
- -----------------------------------------
-