home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!decwrl!adobe!usenet
- From: zstern@adobe.com (Zalman Stern)
- Newsgroups: comp.arch
- Subject: Re: 32 => 64 Transition
- Message-ID: <1992Aug22.002912.10884@adobe.com>
- Date: 22 Aug 92 00:29:12 GMT
- References: <1992Aug22.001234.10390@adobe.com>
- Sender: usenet@adobe.com (USENET NEWS)
- Organization: Adobe Systems Incorporated
- Lines: 28
-
- In article <1992Aug22.001234.10390@adobe.com> zstern@adobe.com (Zalman
- Stern) writes:
- [Complete text of pervious post with no new content.]
- I was selecting a bunch of text to delete because NeXT's NewsGrazer always
- excerpts the entire post when you followup. I accidentally dragged the mouse
- across the post button (while still selecting the text) and sure enough it
- sent it off and deleted the post window. Another of many NeXTStep
- mis-features.
-
- Anyway... What I was going to say is that Alpha is very much a natural 64
- bit machine. It has minimal support for 32 bit operations. The R4000 has
- about equal support (or backwards compatibility for the less charitable) for
- 32 bit and 64 bit.
-
- In an ANSI C universe, the type size_t is used for array indices, the
- arguemnt to malloc. (ptrdiff_t is the type for the difference of two
- pointers. It is should be the same type as size_t only signed.) In the DEC
- model where int is 32 bits and long and pointers are 64 bits, size_t should
- be typedef'ed to unsigned long. That way you can allocate and manipulate
- arrays larger than 2^32 bytes in size. At MIPS, the counter argument to this
- was that little code out there actually conforms to the ANSI standard and
- people use int variables to index arrays all the time. Hence if you want to
- compile dusty decks and have them use arrays bigger than 2^32 bytes, you
- want int to be 64 bits...
- --
- Zalman Stern zalman@adobe.com (415) 962 3824
- Adobe Systems, 1585 Charleston Rd., POB 7900, Mountain View, CA 94039-7900
- "Yeah. Ask 'em if they'll upgrade my shifters too." Bill Watterson
-