vancleef@iastate.edu (Van Cleef Henry H) (06/10/91)
I originally posted this question on comp.unix.ultrix but have gotten no responses. I am using the difference between two gettimeofday() calls to measure real-time performance. What is the number of microseconds added each time the clock ticks. This question applies to 2100, 3100, and 5000 MIPS DECstations running Ultrix 4.*. -- Hank van Cleef vancleef@iastate.edu Iowa State University, Ames. Ia. tmn!vancleef The Union Institute, Cincinnati, Oh.
mellon@nigiri.pa.dec.com (Ted Lemon) (06/11/91)
>From personal experience I know that this is not the best you could wish for >if you are doing real-time measurements. An alternative would be to put a >real-time clock module on the optional VME bus on the DECstation 5000. Gack! That sounds like *serious* overkill! Why not just put one on the TurboChannel? The amount of electronics that you'd be adding to your DS5000 to perform this simple little function is frightening to contemplate... _MelloN_
pilmeyer@utrtsc.dec.com (06/11/91)
In article <MELLON.91Jun10135634@nigiri.pa.dec.com>, mellon@nigiri.pa.dec.com (Ted Lemon) writes... > >Gack! That sounds like *serious* overkill! Why not just put one on >the TurboChannel? The amount of electronics that you'd be adding to >your DS5000 to perform this simple little function is frightening to >contemplate... Yes, I agree. However there is one very good reason to go for a VME clock module instead of a TurboChannel module: There is no clock module for the TurboChannel (at least not when we needed it). -Han
mcculley@alien.enet.dec.com (06/12/91)
In article <1292@sousa.ltn.dec.com>, pilmeyer@utrtsc.dec.com writes... >In article <1991Jun9.233056.12085@news.iastate.edu>, > vancleef@iastate.edu (Van Cleef Henry H) writes... >>I am using the difference between two gettimeofday() calls to measure >>real-time performance. What is the number of microseconds added each >>time the clock ticks. >On the MIPS based DECstations the clock runs at 256 Hz. [...] Henry's original question mentioned that the same question was posted to comp.unix.ultrix but got no responses. For the sake of completeness it probably should be mentioned here that there was a subsequent answer there (from Jeff Mogul) that not only gave the same information as Han's post here, but also added that other clock rates were possible and offered a modified clock driver to implement them. --bruce