CCS :: View topic - Optimization High/Low byte

FAQ

Forum Help

Official CCS Support

Profile Log in to check your private messages Log in

CCS does not monitor this forum on a regular basis.

Please do not post bug reports on this forum. Send them to CCS Technical Support

Optimization High/Low byte

CCS Forum Index -> General CCS C Discussion

View previous topic :: View next topic

Author

Message

RLScott

Joined: 10 Jul 2007
Posts: 465

Optimization High/Low byte

Posted: Fri Feb 25, 2011 11:59 am

In the listing below, I am trying to get the most optimum 16F1824 code for the following function:

Code:

.................... unsigned int16 GetTimer1(void)
.................... {
.................... unsigned int k;
.................... unsigned int16 retGetTimer1;
.................... #byte gt_low=retGetTimer1
.................... #byte gt_high=retGetTimer1+1
....................
.................... k = TMR1H;
*
00BB: MOVF 17,W
00BC: MOVWF 37
.................... gt_low = TMR1L;
00BD: MOVF 16,W
00BE: MOVWF 38
.................... if(k != (gt_high=TMR1H)) //..high byte changed..
00BF: MOVF 17,W
00C0: MOVWF 39
00C1: SUBWF 37,W
00C2: BTFSC 03.2
00C3: GOTO 0C6
.................... gt_low = TMR1L; //..then re-read low byte
00C4: MOVF 16,W
00C5: MOVWF 38
.................... return retGetTimer1;
00C6: MOVF 38,W
00C7: MOVWF 78
00C8: MOVF 39,W
00C9: MOVWF 79

As you can see this is very optimum. But why was it necessary for me to resort to the #byte directives to get it? I was hoping to get the same sort of code from the following:

Code:

.................... unsigned int16 GetTimer1(void)
.................... {
.................... unsigned int8 k,m,n;
.................... k = TMR1H;
*
00BB: MOVF 17,W
00BC: MOVWF 37
.................... m = TMR1L;
00BD: MOVF 16,W
00BE: MOVWF 38
.................... //n = TMR1H;
.................... if(k != (n=TMR1H)) //..high byte changed..
00BF: MOVF 17,W
00C0: MOVWF 39
00C1: SUBWF 37,W
00C2: BTFSC 03.2
00C3: GOTO 0C6
.................... m = TMR1L; //..then re-read low byte
00C4: MOVF 16,W
00C5: MOVWF 38
.................... return ((unsigned int16)n<<8) + m;
00C6: CLRF 3B
00C7: MOVF 39,W
00C8: MOVWF 3A
00C9: MOVWF 3B
00CA: CLRF 3A
00CB: MOVF 38,W
00CC: ADDWF 3A,W
00CD: MOVWF 78
00CE: MOVLW 00
00CF: ADDWFC 3B,W
00D0: MOVWF 7A
00D1: MOVF 7A,W
00D2: MOVWF 79

As you can see, it is nowhere near as good as the first implementation. Is there some way I can coax the optimization of the first listing from an implementation that did not resort to such compiler-dependent constructs as #byte and and assumed endian order? (My build options are set for maximum optimization.)
_________________
Robert Scott
Real-Time Specialties
Embedded Systems Consulting

PCM programmer

Joined: 06 Sep 2003
Posts: 21708

Posted: Fri Feb 25, 2011 1:07 pm

You can use make16() to do it efficiently:

Code:

.................... return(make16(n,m));
000E: MOVF n,W
000F: MOVWF @7A
0010: MOVF m,W
0011: MOVWF @78
0012: MOVF n,W
0013: MOVWF @79

Code:

#include <16F1824.H>
#fuses XT, NOWDT, NOPROTECT, BROWNOUT, PUT //, NOLVP
#use delay(clock=4000000)

#byte TMR1H = 0x17
#byte TMR1L = 0x16

unsigned int16 GetTimer1(void)
{
unsigned int8 k,m,n;
k = TMR1H;
m = TMR1L;
//n = TMR1H;
if(k != (n=TMR1H)) //..high byte changed..
m = TMR1L; //..then re-read low byte

return(make16(n,m));
}

//==========================================
void main()
{
int16 result;

result = GetTimer1();

while(1);
}

RLScott

Joined: 10 Jul 2007
Posts: 465

Posted: Fri Feb 25, 2011 1:45 pm

PCM programmer wrote:

You can use make16() to do it efficiently:

Thanks. That is neater than my solution.
_________________
Robert Scott
Real-Time Specialties
Embedded Systems Consulting

Display posts from previous:

	CCS Forum Index -> General CCS C Discussion	All times are GMT - 6 Hours
Page 1 of 1

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum