Dean Hedin wrote: >I am surprized that Mach under Windows could out perform EMC in steps/sec >since EMC is built on a realtime kernel. > >I presume it is therefore that it is the "quality of steps" that EMC is >better at? In otherowrds EMC produces more accurate and precise steps. > > I haven't put Mach on a scope, so I can't comment on the "quality of steps".
I suspect the main difference is that the Linux systems continues to run normally while EMC2 is operating, which has not been my experience with Mach. When I tried it out on my Athlon 1800, I found that the system slowed to a crawl. System timers were totally screwed up (things like tooltips popping up after 30 seconds instead of 2 seconds), I couldn't run WinAmp - it would skip. I've seen discussions about turning off all sorts of services, tweaking Windows this way and that, etc. to get the best performance out of Mach. I chuckled recently when someone on the Gecko or CCED lists said "I was even able to browse the web when Mach was running, on a P4-3.4GHz". I don't think it's the control of the timer interrupts that makes them different (as Peter Homann suggested), it's the fact that EMC is still sitting on top of a multitasking OS which is still capable of multitasking even while it's running a machine. You could run a web server, play a game, write a thesis, and listen to music while EMC2 is still happily cutting away. All this can be done on a P3-800, without affecting the machine operation (though there will be some reduction in apparent responsiveness on a slower PC). - Steve ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users