+1 use the HPV.
It's an unaccelerated dumb framebuffer (as opposed to the sophisticated accelerated graphics subsystem the T-IV card is) but it's got a 40MHz 128-bit bus leading directly into it's VRAM.
For 2d blitting performance, (almost) no acceleration will ever make up for the difference between a 40MHz 128-bit bus and a 20MHz 32-bit bus. It's the difference between 75MB/s and 600MB/s (approximate absolute peak bandwidth)
It's an unaccelerated dumb framebuffer (as opposed to the sophisticated accelerated graphics subsystem the T-IV card is) but it's got a 40MHz 128-bit bus leading directly into it's VRAM.
For 2d blitting performance, (almost) no acceleration will ever make up for the difference between a 40MHz 128-bit bus and a 20MHz 32-bit bus. It's the difference between 75MB/s and 600MB/s (approximate absolute peak bandwidth)