paul@ppgbms (Paul Evan Matz) (05/05/89)
I was wondering if someone out there might have some idea about what one might hope to expect for interrupt latency behavior on a 386i. We have a video frame grabber we would like to use in a 386i. It has some fairly demanding requirements; It definely does not work with the DOS application that was shipped with it. An interrupt occurs at a 60 Hz rate, which signals the beginning of a 2 millisecond window of time that the board's control registers are accessible. The amount of uninterrupted time then required is estimated, by the board's vendor, to be around 100 microseconds. So, as long as the driver gets about 100 microseconds of execution time sometime within 2 milliseconds after the interrupt, we might be in business. I guess the question is, are these timing characteristics beyond what can be attained by a Unix device driver? I expect we havn't got a chance w/ DOS under Unix, although I hear there may be some better DOS application performance in 4.0.2. Is there anyone up at Billerica that's got an intuitive guess? Thanks (wishful thinking) Regards, Paul Matz PPG Biomedical Systems One Campus Drive Pleasantville, NY. 10570 914-741-4685 path ppgbms!moe!paul@philabs.philips.com