Re: win95

Arvin Meyer (onsite@esinet.net)
Fri, 23 May 1997 11:39:04 -0400


I ran some tests last night and this morning. The 200MHz P5 running under
DESgui 1.2 ran all night (with nothing else running except the blank screen
saver) and got 2 different time sets:

2126/2127secs on 2^30 and
2155/2156secs on 2^30

I was asleep, and nothing changed at all. There was no pattern to which
times ran when. I have no explanation for the difference in times. Even the
power supply on this machine is filtered through a UPC.

This morning I ran another test to see if running 2 instances of the
program would increase the total number of calculations. It did not, on my
single processor. The overall performance degrade was approximately 2%.
That is, it took slightly more than twice the time to run each. BUT, I'd be
interested in knowing what similar tests would produce on a dual processor
machines, especially with machines running NT or other OS's capable of
multithreading.

Arvin Meyer
onsite@esinet.net

----------
> From: Delphin <gt1951a@prism.gatech.edu>
> To: deschall@gatekeeper.megasoft.com
> Subject: Re: win95
> Date: Friday, May 23, 1997 9:50 AM
>
> Jason Gmoser wrote:
> >
> > On Thu, 22 May 1997 23:04:25 -0400, you wrote:
> >
> > >I have a P150 w/ win95 and can do 2^30 keys in <2900 with no other
apps
> > >running...
> >
> > When in Linux my P200 can do 2^30 key pairs in 2040 seconds ('95 gets
2050 or
> > so)... :) It beats the performance of the bitslice code on a 233MHz
alpha that
> > I have access to... The alpha can do 600K k/s. Crazy, no?
> >
> that's fairly fast...my roommates P200 does 2?^30 in about 2300 in
> win95...I'm surprised that Linux really isn't that much faster than
> win95...
> --
> Adam H. Hirschel
> Georgia Institute of Technology, Atlanta Georgia, 30332
> Internet: gt1951a@prism.gatech.edu Campus Phone #1627
> "The sleep of reason breeds monsters" -Goya