Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!gem.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!haven!purdue!mentor.cc.purdue.edu!ajz From: ajz@mentor.cc.purdue.edu (T. Tim Hsu) Newsgroups: comp.sys.intel Subject: Re: 80686 workstation Message-ID: <5024@mentor.cc.purdue.edu> Date: 9 Nov 89 08:29:17 GMT References: <5002@mentor.cc.purdue.edu> Reply-To: ajz@mentor.cc.purdue.edu (T. Tim Hsu) Organization: Purdue University Lines: 39 About the 80686 chip by Intel.... A summer issue of Spectrum magazine (an IEEE publication) gave the details about the 686 chip, but I'll summerize a bit the publication here. The 686 and the 486 were similtaneous projects with the 686 staff being the more prominent group while the 486 staff given the higher priority. I'm going on the assumption that Intel didn't figure on the 686 staff working as fast as they did. The chip is a 128 bit chip with either 121 or 124 bits being used. It is a floating point chip and it has prebuilt 3D vector graphics capabilities. It runs at only 33 MHz, but it is a RISC chip, it performs one instruction per clock cycle (actually it takes a total of three clock cycles to process and instruction, one to load in the instruction onto a cache, one to load the instruction from the cache to the processor, and one for the processor to execute the instruction, but all three phases are occurring similtaneously). I do believe it has a 32K cache on the chip. The article goes on to tell about how the 686 staff had to design around production, how they had to design around the testing procedures, how the existing software to create wafers had to be redesigned to handle a project of such complexity, and how the 686 staff had to also be one step ahead of the 486 staff in order to use the proper machines (since the 486 staff had priority over the equipment). One example of a limitation was that testing could only handle 127 bits and that 486 deadlines kept them from designing more portions of the chip by hand. Is there a catch to all of this? Yes, just one, the 686 chip is no longer compatable with the older model chips, but then again, UNIX runs on anything. Once more, as a query to more informed readers of the net, I thought I saw a 686 workstation made by MIPS, but I'm not sure, can someone verify this, and/or tell me of workstations that you know contain the 686 chip? The 686 seems be a good workstation for intensive graphics based applications. -- T. Tim Hsu UUCP ...pur-ee!ajz@mentor.cc.purdue.edu ARPA ajz@mentor.cc.purdue.edu FAX 1 317 494 0566 BITNET xajz@PURCCVM