Sun Performance under Sunview

[dupuy at cs.columbia.edu]

> We've also noticed that the SparcStation becomes quite slow when someone
> is doing nothing more than editing a file on the console (slow to the
> point that vi acts like the load average is up around 10 or so when it is
> actually less than one).  This is not observed when several people are
> logged on via ethernet -- only when the console is being used.  We have
> lots of swap space (32 Megs), a small optimized kernel and the "standard"
> 327 Meg shoebox.

At a wall street firm I know, they have seen a reduction in performance
going from Sun-4/110s to Sun-4/60s.  They typically run a number of very
large sunview applications, and have seen thrashing when exposing several
windows.  One possibility we considered was the smaller page size on the
Sun-4/60s (4k vs. 8k on Sun-3s and other Sun-4s).

Their salesperson claims they need a new or faster server (this may be
partially true, since they have ~20 diskless workstations on a Sun-3/280
with XD7053 controllers, but they haven't seen comparable slowdowns on
their Sun-4/110s).  The 4/110s have slightly more memory (16M vs. 12M)
than the 4/60s but that doesn't seem enough to account for the differences
they are seeing.

Could anybody "in the know" explain the motivation for reducing the page
size on the sparcstations?

inet: dupuy at cs.columbia.edu
uucp: ...!rutgers!cs.columbia.edu!dupuy