Discussion in PC Magazine and other pubs at the time was that the 486SX-25
was purely a marketing ploy by INTEL to counter competitive pressure from
the AMD 386SX-40. AMD's 40 MHz 386 matched up fairly well against the new
INTEL 486DX-33 in typical user applications, but at a much lower cost. AMD
was starting to make significant inroads into INTEL sales. INTEL had halted
production of the 386 and didn't have time to develop a competitor to the
AMD 386-40, so took the 486DX-33, disabled the FPU and derated the speed (so
it wouldn't take too much market share from the full-blown DX-33 chip) and
marketed it at an extremely competitive price.
>Although this might seem stupid, there was one very good reason for doing
it:
>they could then claim the 386 as being dead (outdated and slow) so that
people
>would stop buying those AMD-386 for fear of getting frigid.
People stopped buying the AMD 386-40 because the INTEL 486SX-25 was priced
so competitively. I recall walking around a computer show back in that
timeframe looking at motherboards and CPUs. I don't recall exact prices,
but it seems to me that a mobo with an INTEL DX-33 was in the high $200s or
low $300s, depending upon manufacturer. I ended up buying a MOBO with a
surface mounted 486SX-25 for only about $125. It would have cost me nearly
that much or perhaps more for the AMD 386SX-40 at that time. The INTEL
486SX-25 was extremely effective; AMD and Cyrix lost most of their market
share gains and didn't reappear as effective competition until just
recently.
>Why didn't they design a true 486SX right away ? Maybe the yields on the
486DX
>were too low so they figured they could feed the chips that failed the
486DX-25
>tests to a 486SX-16 tester. This is much less likely, tho, given the price
of
>testing chips.
The 486SX-25 may have been DX-33 chips which were marginal, but the story in
the industry press at the time was that they were fully functional inside,
just disabled for marketing reasons. At the same time, INTEL pushed MOBO
manufacturers to include an upgrade socket where you could plug in a 487 to
regain the FPU capabilities. Supposedly, the 487 was a normal 486DX-33 CPU
with an extra pin added to the package so that you could only plug it into
these "upgrade" sockets. Doing so disabled the
486SX-25 chip entirely, and the 487 now took over all the CPU functions.
Whether or not this is all true or not, I couldn't say, but it was widely
believed at the time.
Harry
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/