The other way round. The transistor switching times implies the maximum available clock frequency at a given architecture.


Please, as an educated person, you should use proper abbreviations for SI units, here, MHz (note the capital M and H). It DOES make difference.

CMOS is just a name for a technology, in which NMOS and PMOS transistors are used. It does not imply the properties (mainly dimensions, hence speed) of these transistors in any way. Also, the architecture and particular layout of the chip do matter. Although there are variations of CMOS technology allowing very fast switching times of individual gates, these tend to be prohibitively expensive - you wouldn't probably accept a $100 pricetag on a GHz 8051, there are other ways how to achieve higher performance if needed in a particular application, for possibly lower cost. There are also other factors in question than only the price (e.g. power dissipation and power consumption, but also packaging, need for different board layout and power source etc.etc.)

The sheer clock frequency means in fact almost nothing. As noted here, you can get for example a 12MHz P89LPC9xx '51-derivative, which is in fact 6x faster than the "plain vanilla" 8051 - simply because. Also, as somebody already noted, if there are hardware enhancements (e.g. floating point unit) built in, for certain tasks such a microcontroller performs much much faster than the standard one, at the same "frequency".


Jan Waclawek



PS. Please tell us why do you want to run your '51 faster, and we will suggest you the alternatives.