I was wondering what emphasis is going into the development of these?
This is very early in development of the emulators and I have read talk of not enabling the L2 cache, nor using OpenGL ES (which should run "natively" on the PowerVR, right?) and "only" underclocking to 500mhz getting just sick FPS.
Is the focus of the development of the emus just to get something working as a proof of concept? Are we getting to the point that the concepts work and we are eeking as much efficiency from the code as possible to underclock it and save battery life as possible?
That leads me to another question, which is more battery efficient?
1. Leaving L2 cache disabled (so as not to power up that part of the CPU) and just underclocking the processor as much as possible to get it to run fully.
2. Enabling L2 cache so as to underclock the CPU even further because of more efficient use of instructions to the CPU but use more of the CPU itself.
3. Running as much OpenGL code as possible so as to use the GPU to render the 3D (and 2D for that matter) while underclocking the CPU further still.
Which would be the most efficient? I know fully clocking the CPU, enabling L2 cache, and coding as much OpenGL as possible to run on the GPU is going to give you the most power/performance the platform can have. How is the best way to give the highest battery life on these emus that seem to not even be stretching the Pandora's legs?
This is very early in development of the emulators and I have read talk of not enabling the L2 cache, nor using OpenGL ES (which should run "natively" on the PowerVR, right?) and "only" underclocking to 500mhz getting just sick FPS.
Is the focus of the development of the emus just to get something working as a proof of concept? Are we getting to the point that the concepts work and we are eeking as much efficiency from the code as possible to underclock it and save battery life as possible?
That leads me to another question, which is more battery efficient?
1. Leaving L2 cache disabled (so as not to power up that part of the CPU) and just underclocking the processor as much as possible to get it to run fully.
2. Enabling L2 cache so as to underclock the CPU even further because of more efficient use of instructions to the CPU but use more of the CPU itself.
3. Running as much OpenGL code as possible so as to use the GPU to render the 3D (and 2D for that matter) while underclocking the CPU further still.
Which would be the most efficient? I know fully clocking the CPU, enabling L2 cache, and coding as much OpenGL as possible to run on the GPU is going to give you the most power/performance the platform can have. How is the best way to give the highest battery life on these emus that seem to not even be stretching the Pandora's legs?