Post by thread:Optimal realization: Array data to another locatio

I'm glad to say: Greetings All. ;-) I have a little question about how to make couple of FSR from existing one ;-) In words of C it looks like the following: for ( i = num_of_elements; i; ) { b[i] = a[i]; i--; } In words of MPASM it looks slightly different... movlw num_of_elements movwf i loop: decf i,W addlw a movwf FSR ; (a+i) movfw INDF ;**1 movwf temp ;**2 movlw (b-a) addwf FSR,F ; (b+i) movfw temp ;**3 movwf INDF ;**4 decfsz i,F goto loop Did someone see the way how to squeeze it down ? WBR Dmitry. PS. It seems that designers of PIC core didn't like the C ;-)

On Wed, 27 Jan 1999 13:42:31 +0300 Dmitry Kiryashov <spam_OUTzewsTakeThisOuTAHA.RU> writes: >In words of C it looks like the following: >for ( i = num_of_elements; i; ) { b[i] = a[i]; i--; } >Did someone see the way how to squeeze it down ? 14-bit PICs have a hard time with arbitrary moves like that. If you locate the array variables carefully at "block" addresses in RAM, then it's possible to change the pointer between them simply by setting and clearing bits in FSR. For example, place a[0] at $40 and b[0] at $60: movlw array_a ;First byte of A movwf FSR movlw num_of_elements movwf i loop movfw INDF ;get a byte of A bsf FSR,5 ;Point to B movwf INDF ;move into B bcf FSR,5 ;Point back to A incf FSR,f decfsz i,f goto loop For the general case I've looked at using the xoring swap trick to swap FSR with a RAM "fsr2" or W and FSR, but it is faster to use a temporary RAM like you did. ___________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com/getjuno.html or call Juno at (800) 654-JUNO [654-5866]

Hi, The best way is to organize your array a and b that will be exactly multiples of 2 away from each other. ex: b - a = 16 movlw num_of_elements movwf i movlw a movwf FSR loop: movf INDF,w bsf FSR,4 movwf INDF bcf FSR,4 incf FSR decfsz i goto loop regards, Reggie Dmitry Kiryashov wrote: {Quote hidden}> > I'm glad to say: Greetings All. ;-) > > I have a little question about how to make couple of FSR > from existing one ;-) > > In words of C it looks like the following: > for ( i = num_of_elements; i; ) { b[i] = a[i]; i--; } > > In words of MPASM it looks slightly different... > > movlw num_of_elements > movwf i > loop: > decf i,W > addlw a > movwf FSR ; (a+i) > movfw INDF ;**1 > movwf temp ;**2 > > movlw (b-a) > addwf FSR,F ; (b+i) > movfw temp ;**3 > movwf INDF ;**4 > > decfsz i,F > goto loop > > Did someone see the way how to squeeze it down ? > > WBR Dmitry. > > PS. It seems that designers of PIC core didn't like the C ;-)

On Wed, 27 Jan 1999, Dmitry Kiryashov wrote: {Quote hidden}> I have a little question about how to make couple of FSR > from existing one ;-) > > In words of C it looks like the following: > for ( i = num_of_elements; i; ) { b[i] = a[i]; i--; } > > In words of MPASM it looks slightly different... > > movlw num_of_elements > movwf i > loop: > decf i,W > addlw a > movwf FSR ; (a+i) > movfw INDF ;**1 > movwf temp ;**2 > > movlw (b-a) > addwf FSR,F ; (b+i) > movfw temp ;**3 > movwf INDF ;**4 > > decfsz i,F > goto loop > > Did someone see the way how to squeeze it down ? you can decrease the loop execution time by one instruction movlw num_of_elements movwf i movlw b-1 ;initialize FSR movwf FSR loop: movlw (a-b)+1 addwf FSR,f movfw INDF ;**1 movwf temp ;**2 movlw (b-a) addwf FSR,F ; (b+i) movfw temp ;**3 movwf INDF ;**4 decfsz i,F goto loop Scott ps. 34

---------- > From: Scott Dattalo <.....sdattaloKILLspam@spam@UNIX.SRI.COM> > To: PICLISTKILLspamMITVMA.MIT.EDU > Subject: Re: Optimal realization: Array data to another location copyingin words of PIC {Quote hidden}> Date: Thursday, January 28, 1999 12:18 PM > > On Wed, 27 Jan 1999, Dmitry Kiryashov wrote: > > > I have a little question about how to make couple of FSR > > from existing one ;-) > > > > movlw num_of_elements > movwf i > movlw b-1 ;initialize FSR > movwf FSR > loop: > movlw (a-b)+1 > addwf FSR,f > > movfw INDF ;**1 > movwf temp ;**2 > > movlw (b-a) > addwf FSR,F ; (b+i) > movfw temp ;**3 > movwf INDF ;**4 > > decfsz i,F > goto loop If you are free to play with the addresses of your variables you can shorten the main loop anytime the base address is not different by three bits. The address management and temp are 6 instructions worth of overheads. Walter Banks