[Prev][Next][Index][Thread]

Speedometer 3.23 - 1.99p & 1.99p4 comparison



Hi,

I was hoping someone on the list might be able to shed some light on this.  I
saw it was mentioned on another post.  I have used Speedometer 3.23 to
benchmark each new release of Executor and while 1.99o and 1.99p were roughly
comparable, there is an across the board LOSS of between 25 and 35% in
1.99p4.  Have not had the time to try p5 yet, but I am concerned by this
because the only thing which DID speed up is graphics which are almost twice
as fast.  This may sound like a good thing, but it probably only is for
games, and almost all of what I do with Executor involves very little
graphics and mostly other CPU work.  Plus, it helps me make my Mac owning
friends mad that I can run their programs too and they can't run mine.

Does anyone have any suggestions what to try, or should I assume there was
something previously wrong with Speedometer and its prior results are not to
be believed.  I am concerend because that is quite a performance penalty on
the newest upgrade.  I have a AT&T Globalyst 200 notebook computer.  12MB
RAM, 486DX2 50 CPU, and Win95, but often I boot from a DOS diskette.  I know
it has a WD90C24 chipset.  Could this be why it is all falling apart?
 Graphics too slow to give the rest "time" to keep up?

Oh, and a long time ago someone mentioned ARDI was given a free copy of NT by
someone with the Microsoft Employee store.  Are there any plans for ever
making a version of Executor for NT?  I ask because what I have learned about
it lately is that is really is the future of MS platforms and also I think it
could be a great meeting point for legacy WinDOS code and Mac, since you
could conceivably code it all as just another subsystem like the POSIX
subsystem & the Win32 & OS/2 subsystems.  That way either you could run some
Mac apps seamless or have the "browser" as a task built right into the
preemptive protected mode model.

Just a thought...

Greg M.


Follow-Ups: