Linked by Thom Holwerda on Tue 27th May 2008 22:11 UTC, submitted by SReilly
Thread beginning with comment 315872
To read all comments associated with this story, please click here.
To read all comments associated with this story, please click here.
I'm afraid this is mostly nonsense - perhaps due to a badly written benchmark program (not uncommon). Actually, the Cyrix FPU needed 4-7 cycles to do a FADD (a typical operation) while the i387 needed 23-34, and the i287 took 70-100 cycles to do the same.
Regarding FSQRT (important in many early 3D games) the numbers are 59-60 for Cyrix, 122-129 for the i387, and 180-186 for the i287.
On integer code, the Cyrix chip was faster than a Pentium, "clock for clock", and thus much faster than a i486 (see datasheets).
Also, while VIA sold the Cyrix designs for a while, the IDT/Centaur-designed C3 and C7 has nothing to do with the Cyrix chips, at least not regaring technical solutions used.
/Best Regards






Member since:
2007-09-06
How has the design of their processors been coming along? I don't have anything against them really, I like their mini-itx boards, however back in the day when I "upgraded" from a 486dx4 100mhz to a Cyrix PR200 (150mhz), I gained some processor cycles with a huge sacrifice to floating point operations. We benchmarked that thing and it basically said the floating point compared to a 80287 coprocessor... around 1998-99.
, I believe that the VIA chips are originally based off the flawed Cyrix processors, does the new VIA versions fare any better?

IIRC, which I may not
Or maybe I just ramble nonsense?