If you're sure that device is powerful enough to to let dosbox run the games in question, and you already played with the cycles, then you're better off asking the guy who ported dosbox.

I looked up the Nokia N800 and found out the CPU is roughly 330 MHz.
To date my rule-of-thumb has been to apply a 10 times factor for CPU to dosbox performance. Very rough approximation from my circumstances, of course.
Assuming that it applies to the Nokia, I would presume that it should be able to run dos apps that need roughly 33 MHz performance.
No idea how much nethack on dos requires.
If it's too much, there is apparently a nethack port to the Nokia N800 at
http://n770galaxy.blogspot.com/
Might work a bit better.

That's some funky math. Processor frequency does not equal processor speed, not even roughly. There's so many more things to consider to "rate" a performance level - architecture (why is a 1,6 Ghz Dual Core usually faster than a 3 Ghz P4?), optimization (mobile processors are optimized for low power usage, not high performance - why aren't there any 3 Ghz P4's in mobiles and PDA's?), surrounding hardware, etc. .

I dind't try DOSBox on a mobile device yet, but considering the fact that DOSBox is quite CPU-intensive, i'd say it would probably take a lot of tweaking to make games run at good speed on such a device. Another thing to sconsider is how keyboard input is handled by the game. I have some DOS applications that use up all my host's CPU power (it's a P4 2.6 Ghz) when cycles are set to "max", and the app is just sitting at a prompt, waiting for input and doing nothing. I'm no coder, but i guess this is connected to how input is handled by the app. Might be a similar effect with the game you're trying to run.

ADDiCT wrote:

That's some funky math. Processor frequency does not equal processor speed, not even roughly. There's so many more things to consider to "rate" a performance level - architecture (why is a 1,6 Ghz Dual Core usually faster than a 3 Ghz P4?), optimization (mobile processors are optimized for low power usage, not high performance - why aren't there any 3 Ghz P4's in mobiles and PDA's?), surrounding hardware, etc. .

...Very rough approximation from my circumstances...

ADDiCT wrote:

I have some DOS applications that use up all my host's CPU power (it's a P4 2.6 Ghz) when cycles are set to "max", and the app is just sitting at a prompt, waiting for input and doing nothing. I'm no coder, but i guess this is connected to how input is handled by the app. Might be a similar effect with the game you're trying to run.

Good point.

Well that's how max works, it max'es out your processor. If it didn't,
then the implementation would be wrong.
Basically the logic is wrong for the "just sits there" because it doesn't,
it executes callbacks, the idle loop etc.

Well, the settings I've got right now are:
frameskip = 0
cyclespeed = 1144

Does that sound about right?

Also, is there anything else I can do in the dosbox.config file to speed up text based rogue games?

Thanks. 😀

Try cycles=500, and as noted above ask the porter how capable dosbox
is on that device, and if such keyboard problems are known. It surely does
not happen on some 3.2ghz x86 pc.

you definitely need a higher frameskip!

Try a frameskip of 3-4.

cpu cycles with that frameskip should be around 800 on a 770 or 1100-1200 on a N800.

use my dosbox.conf as a starting point. it is optimized for 770.

Okay, will do Pupnik.

Just to repeat my first question again though...

There's not an option in Dosbox to kill the keyboard buffer so keys don't tend to repeat themselves like crazy?

Thanks.

EDIT:
Forgot to mention Pupnik/ArnimS that your default Dosbox.conf file for Dosbox .72 had the speed running at MAXCYCLES for my N800 which crashed the engine. That's why I've been trying to tweak the settings a little.

There's not an option in Dosbox to kill the keyboard buffer so keys don't tend to repeat themselves like crazy?

No, because the feature "addkeysthemselveslikecrazy" isn't implemented,
therefore it can't be disabled.