bt455s39@uhccux.uhcc.Hawaii.Edu (Carmen Hardina) (12/13/90)
In article <1990Dec7.145839.2703@mentor.gandalf.ca> dcarr@mentor.gandalf.ca (Dave Carr) writes: >Talk about people unclear on the concept. The 80186(8) were intended for the EMBEDDED >market. We use the 80186 extensively in our comm products. [....] >The only vendor I know that tried one in a PC was Tandy ! For the job it was designed for, >I stand behind the 80186 any day ! [....] SCI used the 80186 as the primary CPU in their SCI 1000 Tower system a few years ago. The system had a backplane and was not a peecee. It ran AT&T UNIX System III quite well. To this day it will smoke a 286 peecee running UNIX in a multiuser environment. carmen P.S. Does anyone have an Intel 310 or 320 they want to get rid of?
chip@chinacat.Unicom.COM (Chip Rosenthal) (12/16/90)
>In article <1990Dec7.145839.2703@mentor.gandalf.ca> dcarr@mentor.gandalf.ca (Dave Carr) writes: >>Talk about people unclear on the concept. The 80186(8) were intended for >>the EMBEDDED market. No, it was not. It was intended for low-end commercial peecees, and the 286 was intended for high-end peecees. At the time of its introduction, the 80186 was under HIMO (high integration microprocessors operation) and the 80286 was under HIPO (h. i. performance o.) >>The only vendor I know that tried one in a PC was Tandy ! They were the first to have an 80186 peecee. Others did it, like Convergent Technologies (remember them?). My opinion is that the 80186 failed upon two counts. First, the on-board peripherals used several interrupts marked `reserved by manufacturer' on the 8086 datasheet. One of these interrupts was absconded by either IBM (for BIOS) or Microsoft (for DOS). This created a nasty conflict. (I don't remember which - I'd have to go back digging into the datasheet.) The second problem was that the 80186 turned out to be the wrong thing for the market. For example, it's DMA wasn't particularly attractive speed-wise, and few peecees use DMA anyway. Many peecees don't even get it right - just ask Adaptec or somebody else who makes a card which does DMA. Second, gate arrays were to come by just a few years later, creating a technology (and later companies such as Chips and Technologies) which negated a lot of the benefit of the 80186 approach. Not to say the 80186 isn't a nifty embedded controller, but as a CPU it never made it. The embedded controller market is the graveyard for failed CPU's. -- Chip Rosenthal 512-482-8260 | We was raising insurance premiums, ma. Unicom Systems Development | We was spreading fear of arson. <chip@chinacat.Unicom.COM> | - Michelle Shocked
martino@logitek.co.uk (Martin O'Nions) (12/16/90)
bt455s39@uhccux.uhcc.Hawaii.Edu (Carmen Hardina) writes: >In article <1990Dec7.145839.2703@mentor.gandalf.ca> dcarr@mentor.gandalf.ca (Dave Carr) writes: >>Talk about people unclear on the concept. The 80186(8) were intended for the EMBEDDED >>market. We use the 80186 extensively in our comm products. >[....] >>The only vendor I know that tried one in a PC was Tandy ! For the job it was designed for, >>I stand behind the 80186 any day ! >[....] >SCI used the 80186 as the primary CPU in their SCI 1000 Tower system a few >years ago. The system had a backplane and was not a peecee. It ran AT&T >UNIX System III quite well. To this day it will smoke a 286 peecee running >UNIX in a multiuser environment. Research Machines (RM) in the UK used the 186 in their original Nimbus machine. It wasn't a "true compatible", although it did use MS-DOS - some of the H/W was non standard. 3Com used the 186 as the basis for the older 3Server range of dedicated LAN servers. These are still in use on a large number of sites running 3+ Share (DOS based server platform), despite the best efforts of 3Com to get people to trade up to the 386!. -- +------------------------------------------------+-----------------------+ | DISCLAIMER: Nothing Is True (except this) | Martin O'Nions | | ...But if you ask for a rise, it's no suprise | Logitek Group Support | | They're giving none away.... | martino@logitek.co.uk | | (Money - Pink Floyd) | 0257 426 644 | +------------------------------------------------+-----------------------+
dcarr@mentor.gandalf.ca (Dave Carr) (12/17/90)
In article <1755@chinacat.Unicom.COM> chip@chinacat.Unicom.COM (Chip Rosenthal) writes: > >The embedded controller market is the graveyard for failed CPU's. > Yeah, except that it represents about 95 % percent of the market for CPUs. The 80x86 family may be a cash cow for Intel, but their bread-and-butter used to be the embedded market and probably still is.
chip@chinacat.Unicom.COM (Chip Rosenthal) (12/18/90)
Jeeesh...I should stop posting at strange hours of the morning. In article <1755@chinacat.Unicom.COM> I garbled: >At the time of its introduction, the 80186 was under HIMO (high integration >microprocessors operation) and the 80286 was under HIPO (h. i. performance o.) I know, it's meaningless trivia, but I should probably get it right. Make that HPMO, not HIPO for the '286. Also, technically speaking, the CPU's preceded the formation of these operations; they were created briefly after product introduction with the 186/286 as the flagship products. Again, the whole point of these inane details is to explain the 186 was indeed intended as a general purpose CPU. -- Chip Rosenthal 512-482-8260 | We was raising insurance premiums, ma. Unicom Systems Development | We was spreading fear of arson. <chip@chinacat.Unicom.COM> | - Michelle Shocked