Yes it is a bit more voltage than most would like to use, but I slapped a big heatsink and fan on mine and it did just fine.

havli wrote:

Cyrix really isn't that great performance-wise... and runs rather hot too. Looking at my test results MII @ 3*83=250 MHz is just barely faster than P233 MMX and same like K6-2 233. MII @ 3*100=300 MHz beats K6-2 300 (66 MHz FSB) by 4% and loses to Pentium Pro 200/256 😈

AMD K6-2 300MHz (3x100):

Cyrix 6x86MII 300MHz (3x100):

BTW not hot at all because it's 180nm. +4% ALU perf over the K6-2. FPU sucks.

Intel Pentium MMX 300MHz (3x100) overclocked:

ALU sucks. FPU shines.

300MHz Cyrix is much, much faster than 200MHz Pentium Pro. This one is overclocked to 266MHz and still slower on ALU.

This is just one cherry-picked benchmark, how about some real life applications? http://hw-museum.cz/article/5/cpu-history-tou … -1995---1999-/1

And btw - 180nm Cyrix is very hard to find. I have never seen one, just bunch of the older ones and those run very hot.

Well, Ziff-Davis CPU / FPU enchmarks are quick and their results are quite reliable. They don't use MMX, SSE, 3DNow! etc.

When I test overclocked CPUs, I use a simple yet very extensive "benchmark". Boot an old Linux / FreeBSD / NetBSD release and compile a generic kernel with 'time make'. Compilers stress ALUs, branch prediction, caches, memory. It takes 3120 seconds, give or take a few, or 52 minutes to build a generic NetBSD 2.1 kernel on a 300MHz Cyrix 6x86MII + FIC VA-503+ (VIA MVP3, 1Mb cache) + PC100 SDRAM 2-2-2. If a CPU doesn't segfault or produce some other error in the process, consider it a successful overclock. Many overclocked Cyrix CPUs can do benchmarks in Windows yet fail in less than 2 minutes of compile time.

180nm PR400 and PR433 Cyrix 6x86MII with gold tops are quite rare, but there are more common mobile 180nm PR333 with gray tops. These are earlier steppings than gold tops. Although they cannot do 300MHz usually no matter what voltage.