Article View: comp.arch
Article #11112Re: COMPAQ PROPOSED SCALABLE I/O ARCHITECTURE
From: Thomas D. Barrett
Date: Wed, 16 Dec 1992 20:59
Date: Wed, 16 Dec 1992 20:59
44 lines
2472 bytes
2472 bytes
Well... I've only read it through once, but my first impression is that it would be a very elegent architecture for '94. Passing packets of data around on smaller data buses isn't much of a new idea, but as a substitute for a wide PC expansion bus... Obviously in the era of the 8080 one would have been laughed off the podium in making a suggestion that intrs and bus reqs could ever be handled in a timely manner. Today, I can't see it happening for '93, but then again chip designers amaze me at times. If we are talking about '94... let's just go all the way and have three pins: Vcc, Gnd, and fiber... OK, so maybe we want one more Gnd :) Really, though... if you are proposing to build in a lots of "smarts" into modem, floppy, lan, etc. type cards then why not build in a fiber-optic interface? Another thought when you were talking about wide-buses versus your 8-bit proposal... someone is bound to bring up the fact that if your 8-bit bus is going to be sooooo fast with GTL, what is the problem with a 32-bit bus using GTL? Of course, that would be PCI (or VL). I think that for palmtops (or in '94 wrist-tops) definitely need lower pin count without cause a big loss in performance... and the hot-socket is definitely the way to go when presented with limited slots... but I'm a little concerned about suggestions to bring such a technology to the desktop. I *do* agree that ISA and EISA has GOT to go... as a guy who has his PC at home stuffed with serial/modem ports, video capture boards, and LAN cards, I don't enjoy juggling the limited number of IRQs, ROM areas, and time slices... Just think about the fiber proposal... I wasn't joking. Later, Tom -- | Tom Barrett (TDBear), Sr. Engineer | tom.barrett@amd.com | AMD PCD / Austin, TX 78741 | v:512-462-6856 / f:512-462-5155 | "No is Yes, and We are All FREE!" | CO made a #2 no-no... PU! | My views may not be shared by the organization of origin
Technical Details
Message-ID:
<1992Dec16.205908.4826@dvorak.amd.com>Path:
arc1-ca.pugleaf.net!sparky!uunet!stanford.edu!apple!amdcad!dvorak.amd.com!tdbearReferences:
<1992Dec15.171554.2781@twisto.eng.hou.compaq.com>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!stanford.edu!apple!amdcad!dvorak.amd.com!tdbear From: tdbear@dvorak.amd.com (Thomas D. Barrett) Newsgroups: comp.arch Subject: Re: COMPAQ PROPOSED SCALABLE I/O ARCHITECTURE Keywords: I/O, Point-to-Point, High Performance, Low Cost, Bus Message-ID: <1992Dec16.205908.4826@dvorak.amd.com> Date: 16 Dec 92 20:59:08 GMT References: <1992Dec15.171554.2781@twisto.eng.hou.compaq.com> Organization: Advanced Micro Devices, Inc.; Austin, Texas Lines: 44