LAVDRIVER, SYSTEM_MGMT,  Load-Averaging Device Driver

    Here is a VMS driver you can load which will keep track of the
load average in a way you can retrieve it. The algorithm used by it is
not necessarily the same as LAT, but closer to TOPS. To read the
load average, $ASSIGN or open a channel to the device _LAV0: and
read 36 bytes from it. The first 12 bytes are three "F"-type floats
which are the 1, 5, and 15 minute load averages respectively. The
next 12 bytes are the 1, 5, and 15 minute average blocking priority (if
your process is below the average blocking priority you aren't going to
get much CPU, if above it you will get a lot). The last three are the
1, 5, and 15 minute averages of the longest I/O queue length on any disk,
a decent indication of how backed up your I/O system is.

                                                        Kenneth Adelman
                                                        TGV