Game Capsule (NETMF Game Engine)

Just an initial poke at who’d be interested in seeing a game engine for NETMF. I’ve had experience creating game systems before with Direct X and have already done some tinkering with NETMF (like the Demon Solitaire app in Pyxis 2 as well as an Adventure style game and the starts of RPG).

There are some considerations though.

Should it support Cobra and it’s multiple screen sizes without RLP? It would be pretty slow but I know it will support minimal RPGs, Card Games, Adventure Games, etc.

Open/Closed combo? With the original Pyxis I made the OS free but charged $10 for the compiler (compiler is now free by-the-way). How would you feel about a similar approach with a game engine? The engine itself would be free and fully functional but there would also be a commercial RAD environment to make creating the games faster.

And, of course, basic 3D apps for ChipworkX which has RLP open.

I suggest you stick with purely simple game engine. No OS nothing complicated. It would be nice if it is designed around XNA but not exact. Why? Someone already did it and it was too slow!

Maybe some parts can be moved to RLP for performance but you should be able top do it all without RLP. See the game available in the GHI’s graphical demo for example

Haha oh man, 10FPS inside the emulator?! That equates to about 1 frame every 4 seconds in real-world Cobra time.

As for OS stuff, I’m thinking Game Capsule would be separate and stand-alone from Pyxis.

I have seen things run faster on devices than emulator! Do not know why

I’m sure some things do, but graphics aren’t one. I’ve done a large amount of graphics work on NETMF (most of which haven’t made it to Pyxis) and I’ve never seen it be anything but slower on the device. And rightly so, my 8-Core 3GHz machine had [italic]better[/italic] emulate graphics faster than a 72MHz CPU :wink:

I do not k now why, RME runs much faster on our devices that it does on emulator!

Not if it’s a true emulator!

Does the NETMF emulator induce any processing power constraints in order to better simulate a real device?

I am assuming it does since it runs slower then the device. At least this is what I have seen

Maybe try to make it general independent of RLP. Something like:
#if RLP


This way you can run on both with minimal modifications.

An ‘emulator’ only needs to provide the same output given the same input. The way it does so or how long it takes is unimportant. A ‘simulator’ tries to be a realistic model of the target system mimicking its methods (HW and SW). The NETMF emulator is just that, an emulator it has no clue how the actual hardware works, it is only worried about providing the same output as a target system given the same inputs. The speed at which one runs w.r.t. the other has no real world meaning.


Sorry… My defination would be the opposite of yours.

I am currently working on an emulator to verify the results of prior simulation studies.

The simulation was a study of the performance of several new protocols in a large network.

Based upon the favorable results of the simulation, which used discrete simulation techniques, we are now actually building the network nodes,less the actual data plane, to verify the results of the simulation.

@ GHI Mike; yeah if there’s sufficient interest and I create this the way the user interacts with the engine would be the same regardless of which system they were on and RLP would be handled internally.

@ Chris you’ve seen the RPG demo you [italic]know[/italic] there’s a difference in time between the emulator and either the Cobra or CWX


Sorry… My defination would be the opposite of yours.

It is not ‘my’ definition, it is just a fact.


Can we get back on track now?