Missing Multiply Listings

[Charlie U]

Version 3.123, 16F877

I just tried a quick multiply test to see how much code/time it takes. I checked the .lst file, and the assembly code for the multiply is missing. The listing for the multiply is:

.................... c=a*b;
00B2: MOVF 20,W
00B3: MOVWF 25
00B4: MOVF 21,W
00B5: MOVWF 26
00B6: GOTO 004
00B7: MOVF 78,W
00B8: CLRF 7A
00B9: MOVWF 77
00BA: BTFSC 77.7
00BB: DECF 7A,F
00BC: MOVWF 22
00BD: MOVF 7A,W
00BE: MOVWF 23

a and b are signed ints, and c is a signed long.

I reviewed the listing and there is nothing for address 0004, just a big hole. I checked the .hex file and there is code in the "hole", but I haven't tried to disassemble it yet. The program is small (see below) and it does not go over a page boundary. I tried this with 2.374 and the multiply is listed just after the c=a*b; even though the addresses are way out of order.

Has anyone else noticed this?

main .c file:

#include "x_tst1.h"

signed int a,b;
signed long c;

void main() {

setup_adc_ports(NO_ANALOGS);
setup_adc(ADC_CLOCK_DIV_2);
setup_psp(PSP_DISABLED);
setup_spi(FALSE);
setup_counters(RTCC_INTERNAL,WDT_18MS);
setup_timer_1(T1_DISABLED);
setup_timer_2(T2_DISABLED,0,1);
enable_interrupts(INT_TIMER1);
enable_interrupts(global);
a=0x11;
b=0x11;
c=a*b;
printf("c=\%4lx", c);
while (1)
{
;
}
}

include x_tst1.h:

#include <16F877.h>
#device *=16
#use delay(clock=20000000)
#use rs232(baud=9600,parity=N,xmit=PIN_C6,rcv=PIN_C7)
#fuses HS,PUT,WDT
___________________________
This message was ported from CCS's old forum
Original Post ID: 8717

[PCM programmer]

:=Version 3.123, 16F877
:=
:=I just tried a quick multiply test to see how much code/time it takes. I checked the .lst file, and the assembly code for the multiply is missing.
-------------------------------------------------------

Whenever CCS does "optimization improvements", I always cringe.
They always introduce far more bugs than they fix.
Here are some examples from their old versions pages:

2.711 Optimization improvments
2.712 Various bugs in 2.711 fixed
-------------
2.719 Optimization updates
2.722 Optimization bug fixed
------------
3.048 PCW optimization improved
3.048 Some I2C glitches fixed
3.049 A floating point math error is fixed
3.050 Fixed an optimization bug affecting large programs with complex math
3.053 An optimization bug is fixed. This affected some built in functions like math
3.054 Reduced PIC18 optimization to eliminate some user
reported problems while we investigate
---------------

3.093 Linker algorithm has been optimized
3.094 PCB was broke (version 3.093) and is now back
---------------

3.097 Optimization improvments
3.099 Fixed optimization problems (#opt 7 removes new opt, #opt 8 uses it)
-------------------
3.101 More optimization improvments
3.103 Optimization bug fix
3.104 Another optimization bug fix

=============================================

I think we have a pattern here.
___________________________
This message was ported from CCS's old forum
Original Post ID: 8721

[Dale Herman]

Use MPLAB to look at the code.
Some library or built-in sequences don't show up in listing file, but they do in MPLAB.
___________________________
This message was ported from CCS's old forum
Original Post ID: 8739

[Charlie U]

:=:=Version 3.123, 16F877
:=:=
:=:=I just tried a quick multiply test to see how much code/time it takes. I checked the .lst file, and the assembly code for the multiply is missing.
:=-------------------------------------------------------
:=
:=Whenever CCS does "optimization improvements", I always cringe.
:=They always introduce far more bugs than they fix.
:=Here are some examples from their old versions pages:
:=
:=2.711 Optimization improvments
:=2.712 Various bugs in 2.711 fixed
:=-------------
:=2.719 Optimization updates
:=2.722 Optimization bug fixed
:=------------
:=3.048 PCW optimization improved
:=3.048 Some I2C glitches fixed
:=3.049 A floating point math error is fixed
:=3.050 Fixed an optimization bug affecting large programs with complex math
:=3.053 An optimization bug is fixed. This affected some built in functions like math
:=3.054 Reduced PIC18 optimization to eliminate some user
:=reported problems while we investigate
:=---------------
:=
:=3.093 Linker algorithm has been optimized
:=3.094 PCB was broke (version 3.093) and is now back
:=---------------
:=
:=3.097 Optimization improvments
:=3.099 Fixed optimization problems (#opt 7 removes new opt, #opt 8 uses it)
:=-------------------
:=3.101 More optimization improvments
:=3.103 Optimization bug fix
:=3.104 Another optimization bug fix
:=
:==============================================
:=
:=I think we have a pattern here.
:=

I emailed my concern to CCS and they replied that many people were confused by the order the compiler used for library routines such as the multiply. They are placed just after the #device directive from the device .h file. If the (default!)#nolist directive is commented out, the library routines will be included in the listing. I have verified this in 3.123.

Thanks for the replies and help.
___________________________
This message was ported from CCS's old forum
Original Post ID: 8771