home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.arch:10913 comp.lang.misc:3786
- Newsgroups: comp.arch,comp.lang.misc
- Path: sparky!uunet!kithrup!sef
- From: sef@kithrup.COM (Sean Eric Fagan)
- Subject: Re: how to advocate new software/hardware features (Re: Hardware Support for Numeric Algorithms)
- Organization: Kithrup Enterprises, Ltd.
- Date: Thu, 19 Nov 1992 11:12:33 GMT
- Message-ID: <1992Nov19.111233.3108@kithrup.COM>
- References: <martin.721995695@bert> <TMB.92Nov17181419@arolla.idiap.ch> <id.RT2V.U4F@ferranti.com>
- Lines: 23
-
- In article <id.RT2V.U4F@ferranti.com> peter@ferranti.com (peter da silva) writes:
- >You can bet that 64-bit computers will have int=32, long=64, and
- >long-long=128.
-
- 64-bit computers have:
-
- a) int=32, long=32, long long=64 (ELXSI)
- b) int=64, long=64, long long=syntax error (Cray)
- c) int=32, long=64, long long=syntax error (alpha, I think)
-
- (Note that, for the current and upcoming 64-bit microprocessor-based
- computers from the major companies [DEC and MIPS at least, presumably Sun]
- will probably end up having various options for type length, either on the
- command line, or via a header file.)
-
- And throw in combinations of pointers being 32 and 64 bits (or 24 or 32
- bits, on the Cray)...
-
- --
- Sean Eric Fagan | "Time and space is a bitch!"
- sef@kithrup.COM | -- Gooshie (Dennis Wolfberg),
- -----------------+ "Quantum Leap: Killin' Time"
- Any opinions expressed are my own, and generally unpopular with others.
-