Originally Posted by Mike, K8LH

3v0,

May I see the BoostC and MCC18 instructions you used please (sorry for off-topic)?

Mike

Please note that I feel this needs further study. This is the first boostC program I ever compiled. I am not saying it is a bad compiler. I still need to work through each instrucion to see what is going on but it looks like BoostC is testing the bit and setting or clearing it based on the test. What you would do if the bit toggle instruction did not exist.

BoostC

Code:

23:                  while(1)
  004C    D7F4     BRA 0x36
24:                  {
25:                    lata.7 = ! lata.7;
  0036    6A01     CLRF 0x1, ACCESS
  0038    AE89     BTFSS 0xf89, 0x7, ACCESS
  003A    2A01     INCF 0x1, F, ACCESS
  003C    B001     BTFSC 0x1, 0, ACCESS
  003E    8E89     BSF 0xf89, 0x7, ACCESS
  0040    A001     BTFSS 0x1, 0, ACCESS
  0042    9E89     BCF 0xf89, 0x7, ACCESS
26:                    delay_ms(500);
  0044    0EF4     MOVLW 0xf4
  0046    6E01     MOVWF 0x1, ACCESS
  0048    EC02     CALL 0x4, 0
  004A    F000     NOP

MCC18

Code:

17:                  while(1)
  012A    D7F9     BRA 0x11e
18:                  {
19:                    LATAbits.LATA7 = ! LATAbits.LATA7;
  011E    7E89     BTG 0xf89, 0x7, ACCESS
20:                    Delay10KTCYx(50);
  0120    0E32     MOVLW 0x32
  0122    6EE6     MOVWF 0xfe6, ACCESS
  0124    EC65     CALL 0xca, 0
  0126    F000     NOP
  0128    52E5     MOVF 0xfe5, F, ACCESS

I was working on an article for JPUG when I ran into this. I was attempting to show how procedural languages generated similar code. In this case Swordfish and C18. I was curious what BoostC would look like.

LATA.7 is undefined so I used PORTA

Swordfish LATA.7 is undefined so I used PORTA

Code:

7:                 While True
8:                    Toggle(PORTA.7)
  0058    7E89     BTG 0xf89, 0x7, ACCESS
  005A    9E92     BCF 0xf92, 0x7, ACCESS
9:                    DelayMS(500)
  005C    0E01     MOVLW 0x1
  005E    6E05     MOVWF 0x5, ACCESS
  0060    0EF4     MOVLW 0xf4
  0062    DFD3     RCALL 0xa
  0064    D7F9     BRA 0x58

Originally Posted by jay543_uk

can you do the debugging in ccs or do you transfer it over to mplab and do it that way

You do not need to transfer files. CCS integrates with MPLAB. Use MPLAB project wizzard to create project using CCS compiler. After that you can build debug and run from inside MPLAB.

CCS sell its own IDE but it is more money and I prefer to a single IDE for all compiler. MPLAB.

Originally Posted by 3v0

Please note that I feel this needs further study. This is the first boostC program I ever compiled. I am not saying it is a bad compiler. I still need to work through each instrucion to see what is going on but it looks like BoostC is testing the bit and setting or clearing it based on the test. What you would do if the bit toggle instruction did not exist.

Thank you for the example. I'm very impressed that you look at the generated assembly code.

BoostC seems to be missing a toggle_bit() function to compliment their set_bit() and clear_bit() functions. This is one of those situations where having a good assembler background comes in handy as you test different C instruction constructs to find the one that generates the assembler code you're lookin' for (grin).

Anyway, you'll find the following BoostC instruction construct will generate the correct or preferred single word btg lata,7 assembler instruction;

Mike

Code:

  lata.7 ^= 1;

3v0, I think that '!' can't be used as a bitwise operator with BoostC, hence the larger code.


Mike, I wrote my functions that do bitwise operations for Hi-Tech PICC - that did not (and probably still does not) have built-in functions for the task - and they do work with BoostC too.

Code:

#define bitset(var,bitno) var|=(1<<bitno)
#define bitclr(var,bitno) var&=~(1<<bitno)
#define bittoggle(var,bitno) var^=(1<<bitno)

I am not quite sure where you found this but it is incorrect.

Originally Posted by jhanus

In the HiTech, C18 and IAR you can do the following:

RA0=1;
For the C18 you will have to define RA0 like:
#define RA0 PORTAbits.RH0
However there is no way to do such an assignment in the CCS compiler their you have to call a internal...

Peter Anderson has shown how to do this in his execlent tutorials.

Code:

#byte PORTA      = 0xF80

// ------- PORTA Bits -----------------------------------

#bit RA6 	= PORTA.6
#bit RA5 	= PORTA.5
#bit RA4 	= PORTA.4
#bit RA3 	= PORTA.3
#bit RA2 	= PORTA.2
#bit RA1 	= PORTA.1
#bit RA0 	= PORTA.0

As a rule one uses uC and compiler you are asked to work with. To that end I try to be language and compiler neutral.

Microchip is pushing Hi-Tech PICC

I would not read too much into this type of partnership. Both see it as a plus at a very low cost. A marriage of convenience with little commitment.

The CCS demo compiler also ships with MPLAB. I am not sure which parts it supports.

On the other hand the fact that Microchip is releasing their own compilers speaks volumes. They are pushing their own (where they exist).

This is a link to his site. He sells the tutorials for about $50.

Peter H. Anderson - Embedded Processor Control
PC Parallel Port, BasicX, BASIC Stamp®, 68HC11, PIC®, PICAXE, Data Acquisition and Control

It is not suprising

Originally Posted by jhanus

Here, under Ports and Registers, altough it does says CCS have introduced a function in later versions which makes this CCS port assignment unnecessary.
That is wrong too.

I totally agree, but then again I must ask myself if they are all of similar quality how come there exists products which price is tenfold of other.

There must be something. I simply don't want to learn one and after 6 months it struck me, ohhh maybe that wasn't the best choice.

If you were a hobby type you could learn one and stick with it, as an EE you will have to use the compiler the project is using. Most often not your choice.

The price of a compiler has more to do with the nature of the company then quality of the compiler. If you have a small shop with a few programmers you can sell a compiler for much less then if you have a CEO, CFO, HR director and all the departments associated with them. The overhead in some companies is huge.

C18 is a $500 a seat and 5K$ for a site license. The student version lacks some optimization but is free. Given the size and speed of 18F memory chips this is not a limitation you (a student) will bump into soon.

What classes related to programing and embedded systems do you plan to take?

EDIT: What you do not want in a compiler is bad code generation. With a few #defines and macros you can work with most any of them.

I like #inline and #separate compiler directives. CCS has them but they are not unique to CCS. These are used just prior to a function. The #inline forces the compiler to include the code for the function each time the function is used. It avoids the overhead of a function call.

A #inline function can often be used in place of a macro. A function is easier to debug unless the toolchain show the code generated by the macro.

Does BoostC have #inline or similar ?

Originally Posted by 3v0

What classes related to programing and embedded systems do you plan to take?

I just finished third year, orientation automatisation, only class which covered MCU was (in translation) Computer structure, it covered basics of MCU(specifically outdated 16F84A), there was very little interest in my collage(for anything additional), so I covered every function of 16F family(on 16F87xA) alone in assembler.

Now, I'm going to 18F and I need to choose proper compiler, obviously I can't decide should I go PICC or CSSC, probably I'll go with CSSC for more code, tutorials and books to learn from as BeeBop emphasized.