Performance problems with new DOSEMU server

Performance problems with new DOSEMU server

[Danilo Godec]

Hi,

I've hit new problems.The 'old' server, running DOSEMU for multiple 
'telnet' users (there can be 80+ simultaneous DOS sessions) is a 1GHz 
Pentium III based machine with 512 MB RAM....With that much users it was 
getting slow - linux load was hovering at ~8 all the time.

The 'new' server is a 2,83GHz quad core Xeon with 1GB RAM. It should be 
faster then the old one even for a single user, let-alone multiple users 
(theoretically, 4 DOS instances should be able to run in parallel 
without much influence to each other - one for each core, right?).

But test users report the OLD server is about 50 times (!!!) faster then 
the new one in some parts of their applications...I did not yet receive 
instructions to replicate what exactly they are doing so I could do 
tests on my own, but I'd like to do *something* in the meantime.

I tried to get some DOS benchmarks to run on each server, but I cant get 
any to run properly on both server - they either crash, lock-up or 
report non-sense results.

What could I use to compare the 'real life' speed of a DOS instance on 
both servers? For example - if it was linux, I'd compile the kernel and 
measure the time it took...

On the NEW server, a single DOS session is using 15 %CPU doing nothing, 
just sitting at the prompt. I can't do the same test on the OLD one now, 
because users left about 37 DOS sessions to run - but each get's about 6 
%CPU (using 'top' to monitor CPU usage).

I left most of the 'dosemu.conf' alone, these are the only things I changed:

> $_cpu = "80386"
> $_hogthreshold = (10)
> $_xms = (1024)
> $_ems = (1024)
> $_external_char_set = "cp437"
> $_internal_char_set = "cp437"
> $_layout = "us"
> $_mouse_internal = (off)
> $_lpt1 = "/usr/local/sbin/sendprint"
> $_lpt2 = ""
> $_pktdriver = (off)
> $_vnet = "direct"
> $_netdev = "eth0"
> $_novell_hack = (on)
> $_ipxsupport = (on)


Regards, Danilo

Re: Performance problems with new DOSEMU server

[Frank Cox]

On Sat, 2010-02-06 at 07:36 +0100, Danilo Godec wrote:
> But test users report the OLD server is about 50 times (!!!) faster

Is your new server running i386 or x86_64 Linux?

If it's running x86_64, that's the problem.  Running i386 Linux allows
your CPU to use a “Virtual86” mode, while x86_64 Linux requires that the
whole thing be emulated in software which obviously takes more
horsepower and therefore runs more slowly. 


-- 
MELVILLE THEATRE ~ Melville Sask ~ http://www.melvilletheatre.com

--
To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Performance problems with new DOSEMU server

[Danilo Godec]

On 6.2.2010 7:54, Frank Cox wrote:
> On Sat, 2010-02-06 at 07:36 +0100, Danilo Godec wrote:
>    
>> But test users report the OLD server is about 50 times (!!!) faster
>>      
> Is your new server running i386 or x86_64 Linux?
>
> If it's running x86_64, that's the problem.  Running i386 Linux allows
> your CPU to use a “Virtual86” mode, while x86_64 Linux requires that the
> whole thing be emulated in software which obviously takes more
> horsepower and therefore runs more slowly.

I had a hunch that x86_64 might create problems and since the server is 
not supposed to need more then 1GB of RAM, I decided to use i386:

> # uname -a
> Linux dossrv3 2.6.31.8-0.1-desktop #1 SMP PREEMPT 2009-12-15 23:55:40 
> +0100 i686 i686 i386 GNU/Linux


  Regards, Danilo

--
To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html