sef@kithrup.COM (Sean Eric Fagan) (12/19/90)
In article <3060@crdos1.crd.ge.COM> davidsen@crdos1.crd.ge.com (bill davidsen) writes: > The 486 uses fewer cycles than the 386 for the same instructions. The >Weitek can still be added. The boards are easier to design, smaller, and >have less support logic, and are thus cheaper to build. Yes, but the Cyrix '387 clone uses fewer cycles than the '486's on-chip FPU. But you can only add it in a mode similar to the Weitek (i.e., memory mapped). Which is not used by any currently existing code (although I *would* like to write a '387 emulator for the '386 using the Weitek... 8-)). (Note that that should actually be, "much existing code.") -- Sean Eric Fagan | "I made the universe, but please don't blame me for it; sef@kithrup.COM | I had a bellyache at the time." -----------------+ -- The Turtle (Stephen King, _It_) Any opinions expressed are my own, and generally unpopular with others.
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/19/90)
In article <1990Dec19.051616.3763@kithrup.COM> sef@kithrup.COM (Sean Eric Fagan) writes: | Yes, but the Cyrix '387 clone uses fewer cycles than the '486's on-chip FPU. Can you quote me your source on that, and post the numbers? I have run some benchmarks and see the Cyrix as faster than the "slow" 387, slightly faster than the 387-33 (which is new microcode), and slightly slower than the 486, all at 25 MHz. I believe someone has a board which allows you to run up to four 387/Cyrix/ITT chips memory mapped, and some ventor libraries to support it. Since it was a DOS product I didn't make a note of it. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
sef@kithrup.COM (Sean Eric Fagan) (12/20/90)
In article <3067@crdos1.crd.ge.COM> davidsen@crdos1.crd.ge.com (bill davidsen) writes: > Can you quote me your source on that, and post the numbers? I have run >some benchmarks and see the Cyrix as faster than the "slow" 387, >slightly faster than the 387-33 (which is new microcode), and slightly >slower than the 486, all at 25 MHz. The timings I have are all for the "slow" 387, and for teh '486. Basicly, the Cyrix was at least 8 cycles faster on all operations than the "slow" 387, and up to 10 times faster on some of the transcendental functions. The '486, on the other hand, was not quite *that* fast. For simple operations, the '486 is going to be faster than the current generation of Cyrix (a large part of that is because of communication protocols between the CPU and FPU, mind you). -- Sean Eric Fagan | "I made the universe, but please don't blame me for it; sef@kithrup.COM | I had a bellyache at the time." -----------------+ -- The Turtle (Stephen King, _It_) Any opinions expressed are my own, and generally unpopular with others.