C O D I N G
(C) SERGIUS PUZZLER
- Dad, and Th me de
lat with this garbage?
- Yes, simply Zaksor
his son ...
P. Fuzz.
I'll be brief. If you frequently use a lot here this
construction:
...
M_FLAG1 EQU +1
LD A, FLAG
OR A
CALL NZ, PROGA
...
but want to simplify it and save a little, then I suggest
you the following:
...
M_FLAG1
SCF; (or OR A)
CALL C, PROGA
...
Savings of 2 bytes multiplied
the number of such structures in your
program. - Not so much - you exclaim, - but I've not promised -
will be response. SCF include a carry flag and work CALL C. OR
A turn off the flag - CALL C naturally does not work.
A controlled with a flag like this, Makar:
; ON FLAG
LD A, # 37; SCF - FC = 1
LD (M_FLAG1), A
...
; OFF FLAG
LD A, # B7; OR A - FC = 0
LD (M_FLAG1), A
...
; TOGGLE FLAG
LD A, (M_FLAG1)
XOR # 80
LD (M_FLAG1), A
...
Nothing new okromya XOR # 80, which
switches the seventh bit in the cell at
M_FLAG1 tinkering with SCF in OR A, or vice versa.
If you have many addresses like M_FLAG1,
then, for programming convenience, I offer the following:
...
LD HL, **; address of the selected flag
CALL * _FLG; * - SET / RES / XOR
...
...
SET_FLG; INCLUDE
RES 7, (HL)
RET
RES_FLG; OFF
SET 7, (HL)
RET
XOR_FLG; SWITCH
LD A, (HL)
XOR # 80
LD (HL), A
RET
...
Each procedure call SET / RES / XOR teperecha takes into RAM
6/6/6 bts, and held - 5/5/8 bts. Thus the average We did not
win, but not lost but, hopefully, become easier to program. And
yet another procedure SET / RES / XOR (11 bts), ipolzuemaya for
3 or more flags at the same time, saves.
Now a little about one another. In Miracle # 3
If my memory serves me (yes seems to be
no one) has an article about optimizing your code. Its author,
PSV / Dream Team, led one procedure that requires clarification.
To speed up the cycle of the form:
...
LD BC, COUNT
LOOP
...
cycle body
...
DEC BC; works
LD A, B; enough
OR C; slowly
JR NZ, LOOP
...
He offered to do so:
...
LD BC, COUNT
LOOP
...
cycle body
...
DEC C
JR NZ, LOOP
DEC B
JR NZ, LOOP
...
There is no doubt that this version works
faster, but it is checked (even in the mind) has identified an
incorrect setting counter BC. Let me explain: if we run the
following procedure (idiot actually has nothing to do with):
...
LD HL, 0
LD BC, # 0101, 257 (DEC)
LOOP
INC HL; cycle body
DEC C
JR NZ, LOOP
DEC B
JR NZ, LOOP
...
at the exit of HL we find the value equal to 1, instead of
the expected - 257. Conclusion: after setting the counter BC,
it is necessary to adjust - to increase by 256. And here is the
final version:
...
LD BC, COUNT
INC B; missing team
LOOP
...
cycle body
...
DEC C
JR NZ, LOOP
DEC B
JR NZ, LOOP
...
Still not very tired? :) Yes? No? Press
"Y" for Yes or "N" for No. A better fit
stress. Right marker on the monitor.
See you soon. In Kaa ...
SP.WH!
