home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!iWarp.intel.com|inews!doconnor
- From: doconnor@sedona.intel.com (Dennis O'Connor)
- Newsgroups: comp.arch
- Subject: Re: CISC Microcode (was Re: RISC Mainframe)
- Message-ID: <DOCONNOR.92Jul23141720@potato.sedona.intel.com>
- Date: 23 Jul 92 21:17:20 GMT
- References: <BrM8Gv.E3r@zoo.toronto.edu>
- <ADAMS.92Jul21011202@PDV2.pdv2.fmr.maschinenbau.th-darmstadt.de>
- <Brsx7o.G69@zoo.toronto.edu> <2369@nic.cerf.net>
- Sender: news@inews.intel.com
- Organization: Intel i960(tm) Architecture
- Lines: 14
- In-reply-to: davsmith@nic.cerf.net's message of 22 Jul 92 22:49:41 GMT
-
-
- davsmith@nic.cerf.net (David Smith) writes:
- ] All CPUs I have seen to date (not every CPU by any means - if you know
- ] of counter examples, please post) cannot do asynchronous address
- ] generation. When they request a word of memory they want it *NOW* or
- ] within a cycle or two and will block until it arrives.
-
- In the micrprocessor arena, two of the Intel i960(R) families
- military parts, the MM and MX, do support this. They use a split
- transaction bus that can have some number ( 4 ? ) of outstanding
- transactions active at any one time. But I think the motivation
- for doing this was shared-memory multi-processor architectures.
- --
- Dennis O'Connor doconnor@sedona.intel.com
-