Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!crdgw1!uakari.primate.wisc.edu!caen!ox.com!math.fu-berlin.de!uniol!unido!rwthinf!slcdec!hippo!f1.n6000.z2.fidonet.org!f90.n310.z2.fidonet.org!Alexander_Holy
From: Alexander_Holy@f90.n310.z2.fidonet.org (Alexander Holy)
Newsgroups: comp.windows.ms.programmer
Subject: 486SX
Message-ID: <1032427897@f90.n310.z2.fidonet.org>
Date: 3 May 91 08:41:00 GMT
References: <1024862889@p111.f11.n310.z2.fidonet.org>
Reply-To: Alexander_Holy%f90.n310.z2@hippo.dfv.rwth-aachen.de (Alexander Holy)
Organization: Far Side::Vienna, Austria
Lines: 18
Comment-To: Johannes_Blach@f1.n6000.z2.fidonet.org (Johannes Blach)


 > If windows runs slower in 386 mode than in Standard mode, why are wu
 > still using Wimpdos?

I you think twice, Windows makes no sense at all :-)

Actually, a program that mostly reads a 500k file with _lread runs about 5 to 6 
seconds under plain DOS, about 6 to 7 seconds under Windows in Standard Mode, 
and takes about 9 to 10 seconds in 386Enh Mode. The machine is a 386/20 with 
8MB of RAM, so we can be sure there is no swapping.

386Enh makes sense because of simplified memory managment (no segment moving). 
It makes also sense to write 32-bit programs, which run significantly faster. 
The real problem is DOS itself. The most sensible solution would be to include 
DOS itself into the Windows Kernel :-(

---