1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
|
T7.bas a test program for cycle-counting.
The basic idea: we can find out how the real Atari does things by
changing a register after a certain number of NOP's on a certain
scanline. By aligning a player with the colour change, the horizontal
position can be determined and checked on the emulator.
Second idea: We can use RANDOM as a cycle-timer. Set the 9-bit
random number generator (repeating 511 different values in a sequence),
then generate all 511 values and store them in an array, with
the value as the index and the index as the value.
Then given two different readings of RANDOM at different times,
look up the index of both in the table, and the difference , mod 511
is the number of cycles elapsed between the readings.
Only problem is that we can only read 8 bits of RANDOM, so there's
two possible values for each reading, giving 4 possible offsets.
But only one of these will remain constant if the procedure is
repeated.
The 4 values printed by the program at the bottom are the computed
possible offsets, the one which remains constant is the acutal
offset.
N=# of NOP cycles
P=HPOSP0
C=register to change (0=player 0, 9=COLBK, 10=DMACTL, 11-13 HPOS,SIZE,GRAF)
V=value to change to
PG=player graphics
M=antic IR mode
|
