> Just out of curiosity, if I write something for a DOS4GW 32bit > environment, would I choose the same way for having a timer or are there > better ways for having a timer callback function?
the idea will be the same, the routines to use are different. > For a later experiment project I want to capture parallel port data on a > Pentium 3 which is very fast compared to a parallel port's speed. > I thought about having an ISR capturing the data being called at capture > rate, which should be something around 100kHz to 150kHz unlikely. expect *each* inp()/outp() operation to use O(500ns) there is also a notable delay between applying an input 0/1 voltage, and this voltage available as inp() usually people read until 2 readings are identical. try polling first to get an idea what input rate will be possible. > to get a hard > real time measurement. don't mix 'fast measurement' with 'real time, deterministic measurement'. DOS is anything but hard real time. > Are there any interrupt priorities interfering with this concept? in a DOS environment, the current interrupt will be served until finished. so any incoming interrupt (USB, keyboard, mouse) will defer your input routine until done. > There should be plenty of time on this machine to process data after > capturing from parallel port. yes. with real time linux. Tom _______________________________________________ Freedos-user mailing list Freedos-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-user
