Coprocessor instruction fetch identity
This is a dummy address that is generated by the coprocessor whenever it is loading instructions into its own instruction register. This actually occurs every coprocessor cycle except for the second (IR2) cycle of the MOVE instruction. The three types of instructions are shown below.
MOVE | Move immediate to dest |
---|---|
WAIT | Wait until beam counter is equal to, or greater than. (Keeps coprocessor off of bus until beam position has been reached) |
SKIP | Skip if beam counter is equal to, or greater than. (Skips following MOVE inst. unless beam position has been reached) |
MOVE | WAIT UNTIL | SKIP IF |
---|
Bit | IR1 | IR2 | IR1 | IR2 | IR1 | IR2 |
---|---|---|---|---|---|---|
15 | 0 | RD15 | VP7 | BFD | VP7 | BFD |
14 | 0 | RD14 | VP6 | VE6 | VP6 | VE6 |
13 | 0 | RD13 | VP5 | VE5 | VP5 | VE5 |
12 | 0 | RD12 | VP4 | VE4 | VP4 | VE4 |
11 | 0 | RD11 | VP3 | VE3 | VP3 | VE3 |
10 | 0 | RD10 | VP2 | VE2 | VP2 | VE2 |
09 | 0 | RD09 | VP1 | VE1 | VP1 | VE1 |
08 | DA8 | RD08 | VP0 | VE0 | VP0 | VE0 |
07 | DA7 | RD07 | HP8 | HE8 | HP8 | HE8 |
06 | DA6 | RD06 | HP7 | HE7 | HP7 | HE7 |
05 | DA5 | RD05 | HP6 | HE6 | HP6 | HE6 |
04 | DA4 | RD04 | HP5 | HE5 | HP5 | HE5 |
03 | DA3 | RD03 | HP4 | HE4 | HP4 | HE4 |
02 | DA2 | RD02 | HP3 | HE3 | HP3 | HE3 |
01 | DA1 | RD01 | HP2 | HE2 | HP2 | HE2 |
00 | 0 | RD00 | 1 | 0 | 1 | 1 |
IR1 | First instruction register |
---|---|
IR2 | Second instruction register |
DA | Destination address for MOVE instruction.Fetched during IR1 time,used during IR2 time on RGA bus |
RD | RAM Data moved by MOVE instruction at IR2 time directly from RAM to the address given by the DA field |
VP | Vertical beam position comparison bit |
HP | Horizontal beam position comparison bit |
VE | Enable comparison (mask bit) |
HE | Enable comparison (mask bit) |
Note: BFD = Blitter finished disable. When this bit is true, the blitter finished flag will have no effect on the coprocessor. When this bit is zero the blitter finished flag must be true (in addition to the rest of the bit comparisons) before the coprocessor can exit from it`s wait state, or skip over an instruction. Note that the V7 comparison cannot be masked.
The coprocessor is basically a 2 cycle machine that requests the bus only during odd memory cycles. (4 memory cycles per in)It has priority over the blitter and microprocessor.It has priority over the blitter and microprocessor.There are only three types of instructions, MOVE immediate, There are only three types of instructions, MOVE immediate, WAIT UNTIL, and SKIP IF. All instructions require 2 bus cycles (and two instruction words). Since only the odd bus cycles are requested, 4 memory cycle times are required per instruction. (memory cycles are 280 ns).There are two indirect jump registers COP1LC and COP2LC. There are two indirect jump registers COP1LC and COP2LC. These are 20 bit pointer registers whose contents are used to modify program counter for initialization or jumps.They are transfered to the program counter whenever strobe address They are transfered to the program counter whenever strobe address COPJMP1 or COPJMP2 are written. In addition COP1LC is automatically used at the beginning of each vertical blank time.It is important that one of the jump registers be initialized and it's It is important that one of the jump registers be initialized and it's jump strobe address hit, after power up but before coprocessor DMA is initialized. This insures a determined startup address, and state.