We've had OGD1 working as an unaccelerated non-VGA framebuffer device working for some time now. On and off over the last month, we've been working out the major show-stopping bugs in HQ (the MIPS-like microcontroller we developed to emulate VGA). One of the major development problems was that, unlike the Xilinx tools, the Lattice synthesizer seemed to be unable to infer and properly deal with the relationship between the primary clock that runs most of HQ and the double-rate clock that ran its 3-port register file. We worked around that by duplicating the register file (possibly not actually increasing the amount of logic dedicated to it), running it on the primary clock, and extending the results forwarding mechanism to handle simultaneous read and write-back to the same register. It turns out that with the 2x clock, about the fastest we would have been able to run HQ was about 50MHz. With these changes, we can run it easily at 80MHz. Our target is at least 100MHz, but at the moment, this is acceptable for development purposes. We'll defer further optimization to a later time so that we can concentrate on the details of making VGA work. The first step is do develop just enough HQ microcode and x86 BIOS code that we can trick a PC into thinking it's a real VGA card; the objective is to be able to boot into Linux without a functional display so that we can incrementally develop the code that will make VGA actually put up a working display.
-- Timothy Normand Miller http://www.cse.ohio-state.edu/~millerti Open Graphics Project _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
