Aaron Lehmann wrote:
 
> Not to mention that an emulation of a more modern card would support
transfer rates far in exess of 10mb/s.
 
Uuh... The hardware sets the speed limit. An emulation of a fast card that ultimately goes through a slow card will be slow. An emulation of a slow card that ultimately goes through a fast card could be fast (at least on Windows), depending on the efficiency of the emulating code, of course.
 
I do think that ultimately we'd be better of with our own "virtual" LAN card for which we write a driver (not much to it on Windows). It could cut out a whole layer of code, making for that much better performance. Of course, it wouldn't know anything about the PXE capabilities that Ohso mentions, but I don't really see how that would work under plex86, anyway. Anyway, no reason not to have several approaches, so that users can choose the best one for their circumstances. Choice is a *good thing*.
 
Once I get the (windows?) double-headed video running, I'll be taking a look at an NDIS driver for a virtual LAN card.
 
While I'm here, sorry about the earlier double posting - the date roll-over was late on the mailing list and I thought the first one had lost itself in netland.
 
Cheers, Colin

Reply via email to