Simply run a benchmark and compare.

But i can't compare, can you suggest wich benchmarck?

It seems to me that this is comparing apples and pears. The slowest Pentium III runs at 450Mhz and it is very unlikely that any Pentium of that period at 366Mhz can be as fast or faster.
You could downclock a Pentium III 550Mhz to 366Mhz via FSB 66 but that would defeat the purpose of the comparison.
But if you want to go into benchmarking you can find some suggestions in this document:

https://en.wikichip.org/w/images/5/51/Mobile_ … %2C_1999%29.pdf

You can compare Dixon to Mendocino at similar clock, but with a few percent higher performance due to larger cache. Probably a Dixon 366MHz will beat a P2-350 in most tasks and depending on workload (whether things fit in L2 or not) might approach P2-400, although if workload doesn't fit in 256kB, the slower memory interface (66MHz Dixon vs 100MHz Deschutes) would strongly prefer Deschutes.

Articles at the time stated that the 256KB on-die cache tended to perform better than the 512KB external cache, and obviously it would be an improvement over mendocino. Not as fast as coppermine though because the L2 on those is connected with a wider bus for lower latency.

Use this tool:
http://testcpu.webz.cz/

How about Everest Home Edition for a benchmark test with comparative results included. Or Sisoft Sandra 99?

bakemono wrote:

Articles at the time stated that the 256KB on-die cache tended to perform better than the 512KB external cache, and obviously it would be an improvement over mendocino. Not as fast as coppermine though because the L2 on those is connected with a wider bus for lower latency.

And let's not forget the advantage of SSE instructions in the Pentium III starting with the Katmai.