: > Does anyone have the MIPS and MEGAFLOPS rating of a fast 486 ESA
: > system? WE are constantly arguing here at work about speeds and I was
: > wondering if anyone has benchmarked these things lately.
:
: Sorry, no MIPS or MEGAFLOPS (or SPECmarks) ratings, but rather I do have
: a real-world, number-crunching application. Assuming a 50 MHz 486 is
: twice as fast as a 25 MHz 486 (which is what I tested), then the Sparc-
: station 2 is no more than 10 percent faster than the 486. No doubt
: different results would be obtained for graphic-intensive applications,
: though you'll also see a wide variance among 486 systems, due to the
: different ways they implement video. Even among number-crunching
: applications, it still depends on the application. I have one application
: for which the 25 MHz 486 runs circles around a Sparcstation 2, but I have
: yet to isolate what code fragment the Sparcstation 2 is falling down on.
:
: Note that my benchmarks were done using 32-bit code on a 32-bit operating
: system (OS/2 2.0), for which I find a factor of two improvement in performance
: over their older 16-bit counterparts. Perhaps that's part of the underlying
: misconception that PCs can't hold a candle to Suns; you lose a factor of two
: (when comparing applications that benefit from a 32-bit data path) by
: comparing a computer running a 16-bit application to another running a 32-bit
: application. I'm very happy with my 25 MHz 486, which typically performs at
: the level of a Sun IPC for my number-crunchers.
PC-Magazine recently compared the speed of a 486/50, Sun, Next, Mac, etc and
found that the 486/50 was the fastest of the lot. They conducted benchtests
by timing certain processor intensive operations in wordprocessors, spreadsheets, and desktop publishing using the same programs on all platforms. It is quite
