The conventional wisdom in the days of the x87 math coprocessor was that
you wanted the math co to run at the same speed or faster as the CPU.
Matching SX with SX and DX with DX was more important than speed. If the
software you run does not rely on the 387 coprocessor for math-intensive
calculations, it would never know whether or not an x87 chip was
installed, much less how fast it might be running. If the program does
rely on the math unit for math-intensive calculations, that part of the
operation would be a little slower if the math unit runs at 33 MHz and the
CPU runs at 40 MHz. It would still work, and in most cases, you probably
would not hnotice much of a difference in actual real-world time needed to
perform a given task. If you're modeling the wing surfaces and top-line
profile for the next small corporate jet aircraft the difference might be
critical. If you're writing and testing a program to calculate and keep
track of sales and value-added taxes using a program like MicroFocus
COBOL, it won't make much of a difference.
I knew a fellow who ran a 20 MHz 80387SX chip in a notebook computer with
a 25 MHz 80386SX CPU for several years with no ill effects. If you're
going to run that 80387 one clock rating faster than its designated speed,
make sure it's a good one. In the days of the 80387 chips, Cyrix was
considered to have made the best ones on the market. They claimed, and
had a testing program that proported to prove it, that their 25 MHz 80387
processor performed 37 percent faster than Intel's 80387 processor rated
for the same clock speed. If that 33 MHz 80387 processor is a Cyrix, it
just might perform on a par with an AMD 40 MHz processor anyway. Cyrix
made their original reputation for excellent performance and quality on
the strength of their X87 math coprocessors. That's one of the main
reasons why so many people were so disappointed when the math units in
their 486- and pentium-class processors lagged behind both the Intel and
the AMD math units in similar class processors.
Brent Reynolds
Random Access Internet Shell account
Standard disclaimers apply.
Email: [EMAIL PROTECTED]
To unsubscribe from SURVPC send a message to [EMAIL PROTECTED] with
unsubscribe SURVPC in the body of the message.
Also, trim this footer from any quoted replies.
More info can be found at;
http://www.softcon.com/archives/SURVPC.html
