>> It looks like the binary dump in the latter file is formatted >> differently. When you strip the address and trailing * and wrap >> the lines at 32 instead of 40 bits per line, it starts to look >> more like the earlier dump. > > Yes, except it is complete ;-) > > They are not identical; I'll have a look what the difference is. > Interesting. It turns out that all bit pairs in the AND array are swapped between the two JED files. They are identical otherwise (ignoring the missing lines). I suppose one of the JED file programs has a bug, or alternatively one of the devices inverts all its inputs. The JED bitstream is 48 lines of 40 bits, and a final 8 bits. Each line is 16 pairs of inputs for the AND array, creating a minterm; and then 8 bits for the OR array, saying which sum terms use that minterm. The final 8 bits are to optionally invert the outputs. The inputs are (pin #, pin name, c64 pin name, where does the signal come from): 9 I0 #CAS (VIC) 8 I1 #LORAM (CPU P0) 7 I2 #HIRAM (CPU P1) 6 I3 #CHAREN (CPU P2) 5 I4 #VA14 (CIA2 PA0) 4 I5 A15 (CPU) 3 I6 A14 (CPU) 2 I7 A13 (CPU) 27 I8 A12 (CPU) 26 I9 BA (VIC) 25 I10 #AEC (VIC) 24 I11 R/#W (CPU) 23 I12 #EXROM (cartridge) 22 I13 #GAME (cartridge) 21 I14 VA13 (VIC A5/A13) 20 I15 VA12 (VIC A4/A12) and the outputs are 18 F0 #CASRAM (#CAS on RAM) 17 F1 #BASIC (basic rom #CS) a000..bfff 16 F2 #KERNAL (kernal rom #CS) e000..ffff 15 F3 #CHAROM (char rom #CS) d000..dfff 13 F4 GR/#W (colour RAM #WE) d800..dbff 12 F5 #IO (I/O #CS) d000..dfff 11 F6 #ROML (cartridge #CS) 8000..9fff 10 F7 #ROMH (cartridge #CS) e000..ffff All the equations in pla.txt are correct as far as I can see. There are two minterms computed that aren't used: ras, addr 1101, cpu write cas My guess is that those were used on some earlier revision and never removed. The first is right before the I/O minterms, the other before the CASRAM minterms. Segher Message was sent through the cbm-hackers mailing listReceived on 2012-05-10 19:00:36
Archive generated by hypermail 2.2.0.