home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!cs.utexas.edu!torn!news.ccs.queensu.ca!qucdn!leek
- Organization: Queen's University at Kingston
- Date: Thu, 7 Jan 1993 14:09:17 EST
- From: <LEEK@QUCDN.QueensU.CA>
- Message-ID: <93007.140917LEEK@QUCDN.QueensU.CA>
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: 2.0 UNIT_MICROHZ timer (questions...)
- References: <93003.144651LEEK@QUCDN.QueensU.CA>
- <1993Jan5.110742.11479@rhrk.uni-kl.de> <93005.144922LEEK@QUCDN.QueensU.CA>
- Lines: 35
-
- I got the problem solved. I called Delay() instead of my own
- MicroDelay(Sec,Micro). Under 2.1, the OS came back in < 30uS later as the
- first parameter = 0. Thanks for all that emailed me or posted. The following
- is a list of the actual timing I measured. My routine was based on the RKM
- timer.device example that appeared on Fish disks.
-
- Delay (mS) Measured Delay (mS)
- ---------- -------------------
- 10 11.0
- 20 21.0
- 40 41.0
- 80 81.7
- 100 101.0
-
- Note:
-
- Delay : value passed to UNIT_MICROHZ timer.device
- Measured Delay: Actual ellapse time from setting a bit in the parallel port,
- to making a request for the time delay and coming back from
- the IO operation to resetting a bit in the parallel port.
-
- Measurement done with task priority set to 25 on a 18.3MHz
- 68020 A1000 (NTSC, Eclock =716Khz), OS 2.1 with light load.
- ROM is in 16-bit and running at 7Mhz bus speed.
- Routine coded in Aztec 3.6a C code with modified RKM example.
- Timing done on a Philips Logic Analyzer.
-
- As you can see on my setup, the system timer device is pretty good. The system
- overhead is about 1 to 2 millisecond (include task switches, OS overhead, and
- actual IO port overhead in C).
-
- Thanks again guys/gals !!
-
- K. C. Lee
- Elec. Eng. Grad. Student
-