Dale Martin said: >don't support. We have to think about: >- All of the x86 compatible Intel processors All except the actual i386 (all suffixes) support the 486 instruction set.
>- All of the x86 compatible AMD processors All except the 386* chips support the 486 instruction set. >- All of the x86 compatible Cyrix processors Most except the 386* chips support the 486 instruction set. There may be an odd exception (a '486' chip which isn't) although I doubt it. >- All of the x86 compatible VIA processors All support the 486 instruction set, I believe. >- All of the x86 compatible TransMeta processors All support the 486 instruction set, I believe. >- The National Semiconductor Geode Supports the 486 instruction set, I believe. >- What else? Some chips from companies which are now bankrupt or merged out of existence. :-) The "WinChips" you mentioned support the 486 instruction set. During the 386 cloning period, all the cloners called their chips "386" or "80386". Unique naming by Cyrix, AMD, etc., only really started later, so basically, if it doesn't claim to be a 386 (or earlier), it's something later. (Intel called theirs the 'i386' so that there would be a distinction between the Intel 80386 and everyone else's.) After the 486, Intel always provided a method to determine the CPU type and features available. As far as I can tell, there's no easy programmatic way to tell the difference between a (old, no CPUID) 486 and an (old, no CPUID) 386 without risking an illegal instruction, which bites. :-( (I could be wrong.) The 386 has a few serious defects for multitasking operation (which are by now well documented), all of which were fixed in the 486. I wouldn't recommend using it for anything net-connected and supposed to be secure. Apart from the slow-as-dirt problem, of course (the first ones were slower than 80286s...). As someone else mentioned, i386s are vanishing pretty quickly due to component failure. Cheap or discarded 486s are available by the busload to replace them, anyway... --Nathanael
