Xref: news-dnh.mv.net comp.os.msdos.djgpp:2195
Path: news-dnh.mv.net!mv!news.sprintlink.net!in2.uu.net!noc.near.net!das-news2.harvard.edu!oitnews.harvard.edu!newsfeed.rice.edu!rice!news!sandmann
From: Charles Sandmann <sandmann AT clio DOT rice DOT edu>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: lots of interrupts per sec with DJGPP
Date: Tue, 26 Sep 1995 09:49:06 CDT
Organization: Rice University, Houston, Texas
Lines: 13
References: <DFH8HE DOT 7x AT uns DOT bris DOT ac DOT uk> <306712ca DOT sandmann AT clio DOT rice DOT edu> <DFILwJ DOT Gqq AT uns DOT bris DOT ac DOT uk>
Reply-To: sandmann AT clio DOT rice DOT edu
Nntp-Posting-Host: clio.rice.edu
To: djgpp AT sun DOT soe DOT clarkson DOT edu
Dj-Gateway: from newsgroup comp.os.msdos.djgpp

> I discovered the following when run on a PCI 66Mhz 486:
>           Using standard go32 PM wrapper      Using simple asm wrapper
>    XMS    3000-4000                           70000-75000
>    DPMI   8000-9000                           70000-75000
> Is there any reason why the standard interrupt wrapper routines are so slow?

The only way the go32 wrappers can be so slow is if you chain (and these numbers
are consistent with what I see with chaining).  The simple asm routine 
numbers make sense with only a PM ring change (and no swap to real mode).
The wrappers should only add a few hundred cycles per interrupt.

The reason XMS is slower is the toggling of the A20 line under HIMEM.SYS
is relatively slow.