Sorry this is a bit of a late reply, got a bit sidetracked by life but am here all the same..
I actually have a GCW-Zero and am also the guy that ported rRootage to the GPH Wiz. See: http://dl.openhandhelds.org/cgi-bin/wiz.cgi?0,0,0,0,30,379
The Wiz had a ARM CPU with no FPU to speak of (all floating-point done in software!) clocked at 533mhz but capable of 733mhz overclock in practice.. It did, however, have very basic & slow OpenGLES support. The original PC rRootage used almost-exclusively floating point math and also was designed for a very old, completely-incompatible OpenGL spec, 1.2.
The way I got it working was to completely convert all OpenGL-related rRootage code to heavily-optimized OpenGLES 1.2. Furthermore, to achieve playability on the Wiz, I converted most all math to software-based fixed-point rather than floating point. There were an extremely-few small parts of the math that overflowed 16.16 32-bit fixed point and so I had to retain those few bits as software-implemented floating-point. It was quite playable at 700Mhz for Levels up to 8 (default overclock I instituted for the program).
With the CPU of the GCW-Zero having 1GHz speed as a default, an FPU unit, DDR2 RAM, as well as a much-more advanced GPU (Once drivers become stable) rRootage should assuredly run at full-speed. I have it on my to-do list for early 2014. I have even already implemented re-configurable controls, optimized layout for a 320x240 rotated screen, and other customizable options. So, yeah, rRootage is a sure thing.
Some might now say, "Well, what about NanoGL or *insert crappy OpenGL->OpenGLES wrapper library here* but the truth is they're usually designed with a specific game in mind (think Quake) and are far from bug-free or feature-full and because of the nature of how the OpenGL 1.2 interface is so ancient compared to OpenGLES, they are slow
. I actually went that path with terrible speed, stability, and visual quality until just saying "screw it, just rewrite it properly and take advantage of OpenGLES advantages" and was only then satisfied with the results.
Now, as for Tumiki Fighters, Torus Trooper, etc.. They are all written in the D language, whereas rRootage/Noiz2sa were written in C++. There is no D compiler available for the MIPS architecture. Therefore, those would all require not only a complete rewrite of the source code from the D language to C++, but also, like my rRootage port, a complete rewrite from OpenGL 1.2 to OpenGLES. Not exactly an easy task, and their CPU loads might be too high versus rRootage's.
Only experimentation and programmer-sweat can tell. The chances are much more slim for these ports being completed but I already have thought quite a bit about them and looked at the source code a bit several years ago thinking about all this. So, there's that at least hehe