[oclug] Gods thing

[John E Pearson]

On graphics comparisons:

I was told this morning that graphics cards benchmarks are pretty hard to
do. In fact I was told that they aren't really done because there are too
many unknowns. I did a test that I thought would compare the
Gloria II agaisnt the Gloria III. The III is supposed to be about twice as
fast as the II. The Gloria III was on a 733 MHz box and the Gloria II was
on an 850MHz box. The Gloria II won the race by a little. this means that
my test was inadequate to distinguish between the two cards and that a
bunch of the geometry was being done by the CPU rather than the graphics
card.  I think now that without the guidance of a graphics expert that my
quantitative comparison proposal is pointless.

On X being a memory hog: I was told that both X and windows will grab lots
of memory but give it up to applications as needed. This is consistent
with my experience and the data I posted yesterday. Does anyone know
whether it makes sense to do all this optimization (as discussed by H V-S)
to reduce the amount of memory that X grabs? I think it doesn't make any
sense because I think that X isn't really such a hog.  I don't know what
word to use to describe something that hoards resources but gives them up
willingly to anything that needs them. It probably wouldn't be "hog". I
guess you could find out how much you pay for running X by running a CPU
intensive application with and without X running ? Is that an adequate
test? It seems to me that running ps or top to see how much memory X is
using is essentially meaningless. Isn't the crucial issue how running X
impacts other applications?

jp