multiplication vs shifting

[Ioannis]

No other code, just Variable declaration and END,for a F877:

For BYTE variable:

x=x*25            '40 bytes
x=x<<4+x<<3+x     '43 bytes

For WORD variable:

x=x*25             '43 bytes
x=x<<4+x<<3+x      '49 bytes
x=x<<25            '27 bytes!!!

Ioannis

[Kamikaze47]

whats the point of the last one? shift a word variable 25 times either way and you will just get 0.

i was multiplying a word var and saving into a long var

[Ioannis]

I know 25 shift will not give the result you want.

Just tested to see how much less program space and how faster a shift will be. Useless for you but quite impressive!

Which PIC are you using? Is there any chance to use one with hardware multiplier?

Ioannis

[Darrel Taylor]

To me, the time it takes to execute is more important than the size of the code.
And the fastest way will usually <strike>be the smallest too</strike> not be the smallest.

On a 16F @ 4Mhz with WORD variables ...
<table cellpadding=6 border=1><tr><td align=center>Formula</td><td align=center>Execution Time</td><td align=center>Words per<br>Instance</td><td align=center>Words added<br>to Library</td></tr><tr><td>x = x*25</td><td align=center>230 - 250uS<br>depends on X value</td><td align=center>12</td><td align=center>27</td></tr><tr><td>x = x<<4 + x<<3 + x</td><td align=center>110uS</td><td align=center>34</td><td align=center>13</td></tr><tr><td> T = X << 1
T = T << 1
T = T << 1
X = T << 1 + T + X </td><td align=center>31uS</td></td><td align=center>30</td><td align=center>13<br>adds SHIFTL to library<br>but doesn't use it.</td></tr></table>

I wonder if ASM will help

[Kamikaze47]

Nice work DT.

It's amusing that doing a single shift twice takes less time than a double shift.

I have run into a problem though. Does anyone know why this gives me an incorrect result?

X VAR WORD
T VAR LONG
Y VAR LONG

X=8100

T = X << 1
T = T << 1
T = T << 1
Y = T << 1 + T + X

LCDOUT $FE,$80,"T=",DEC T
LCDOUT $FE,$C0,"Y=",DEC Y

T is 64800 (as expected), but Y is 2516719364 ?!

I'm guessing that its something about long variables that I am missing.

[Darrel Taylor]

Nice find Kamikaze!

There's a small bug in the pbppi18l.mac file.
The SHIFTL?NCN macro should be changed as follows. (3's in red)

SHIFTL?NCN macro Nin, Cin, Nout
    if ((Cin) == 1)
        bcf     STATUS, C
      if ((Nout) == (Nin))
        CHK?RP  Nout
        rlcf    Nout, F
        rlcf    (Nout) + 1, F
        rlcf    (Nout) + 2, F
        rlcf    (Nout) + 3, F
      else
        CHK?RP  Nin
        rlcf    Nin, W
        MOVE?AB Nout
        CHK?RP  Nin
        rlcf    (Nin) + 1, W
        MOVE?AB (Nout) + 1
        CHK?RP  Nin
        rlcf    (Nin) + 2, W
        MOVE?AB (Nout) + 2
        CHK?RP  Nin
        rlcf    (Nin) + 3, W     ; was 2
        MOVE?AB (Nout) + 3
      endif
    else
        MOVE?NN Nin, R0
        movlw   Cin
        L?CALL  SHIFTL
        MOVE?ANN R0, Nout
    endif
    endm
SHIFTL_USED = 1
  endmod

I'll let the guys know.
<br>

[Kamikaze47]

Yep, that fixed it. Thanks DT.

I'm surprised that the bug hasn't been found before now.

Out of curiosity, why did it only affect the shift on the last line, and not the first three?

T = X << 1
T = T << 1
T = T << 1
Y = T << 1 + T + X