Article View: comp.os.os2.programmer
Article #969Re: Does math coprocessor speed up GUI (here: OS/2 presentation mgr)?
From: Stephen M. Lacy
Date: Wed, 26 Aug 1992 10:47
Date: Wed, 26 Aug 1992 10:47
43 lines
2747 bytes
2747 bytes
Excerpts from netnews.comp.os.os2.programmer: 26-Aug-92 Re: Does math
coprocessor s.. Al Dunsmuir@TOROLAB6.VNE (1580)
> Having said that, having a math coprocessor (especially a tightly coupled
> one as implemented in the 486) can speed up OS/2 applications. For example,
> IBM's X-Windows server (PM screen interface for TCP/IP) achieves __dramatic__
> performance gains when a math coprocessor is present.
Why is this so? I thought that X used all integer based math too. The
reason that I'm asking is that I just saw PMX on a machine other than my
own for the first time a couple days ago, and I was really really
impressed. It was not much slower than running right on the console of
one of the DecStations here at work, where on my computer it was
spitting out text at maybe 4800+ baud (which is slow) I was also
trying to figure out where the greatest difference in speed was, here
are the two setups:
My Computer: 16 Mhz '386 10 Meg RAM
Token Ring Card (4 mega-bits per second DTR)
Regular VGA (8-bit)
Friends Computer: 50 Mhz '486 8 Meg RAM
EtherNet Card (10 mega-bits per second DTR)
Trident video card running in VGA mode
My first idea was the video card, since video I/O seemed very slow on my
computer, but then I thought "well, maybe its the network card" What
convinced me that this was not trus is that both of our FTP's from a
local machine (i.e. a machine that we would be running X applications
from) achieve about the same DTR. (DTR=Data Transfer Rate) So, I
thought for a while it was just the speed of the computer, but I don't
think that would make all that much difference in this situation. You
now mention the co-processor, and that it gives a large speedup. Could
the co-processor be where the real speed lies? If getting myself a
cyrix would be a cheap way to have it run at a more reasonable speed,
I'll order one today!
Thanks.
"Now it's over I'm dead and I havn't done anything that I want -- or I'm
still alive and there's nothing I want to do." -They Might Be Giants
Technical Details
Message-ID:
<Qeax6pS00Vot8FvkYs@andrew.cmu.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!sun-barr!cs.utexas.edu!uwm.edu!ogicse!das-news.harvard.edu!cantaloupe.srv.cs.cmu.edu!crabapple.srv.cs.cmu.edu!andrew.cmu.edu!sl31+References:
<17fuahINN1tl@iraul1.ira.uka.de> <19920826.074333.779@almaden.ibm.com>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!sun-barr!cs.utexas.edu!uwm.edu!ogicse!das-news.harvard.edu!cantaloupe.srv.cs.cmu.edu!crabapple.srv.cs.cmu.edu!andrew.cmu.edu!sl31+ From: sl31+@andrew.cmu.edu (Stephen M. Lacy) Newsgroups: comp.os.os2.programmer Subject: Re: Does math coprocessor speed up GUI (here: OS/2 presentation mgr)? Message-ID: <Qeax6pS00Vot8FvkYs@andrew.cmu.edu> Date: 26 Aug 92 10:47:17 GMT Article-I.D.: andrew.Qeax6pS00Vot8FvkYs References: <17fuahINN1tl@iraul1.ira.uka.de> <19920826.074333.779@almaden.ibm.com> Organization: Sophomore, Math/Computer Science, Carnegie Mellon, Pittsburgh, PA Lines: 43 In-Reply-To: <19920826.074333.779@almaden.ibm.com>