EE353 L03 - PIC24 Family.ppt

8/11/2019 EE353 L03 - PIC24 Family.ppt

1/75

1

Introduction To PIC24Family

Sohaib Majzoub

Lecture Notes 3

EE 353Introduction to Microprocessor

Lecture notes by Dr. Bob Reese, ECE Department, Mississippi State UniversityModified with permission by Dr. Sohaib Majzoub, College of Engineering, King Saud University.


2/75

2

Microcontroller (C) vs. Microprocessor

(P) C intended as a singlechip solution, P requires

externalsupport chips(memory, interface)

C has on-chip non-volatile memoryfor programstorage, P does not.

C has moreinterfacefunctions on-chip(serial

interfaces, analog-to-digital conversion, timers, etc.)than P

C does nothave virtual memorysupport (i.e., couldnot run Linux), while P does.

General purpose Psare typically higherperformance(clock speed, data width, instruction set,cache) than Cs

Division between Ps and Cs becomingincreasingly blurred


3/75

3

Microchip PIC24 Family C

Features Comments

Instruction width 24 bits

On-chip program memory (non-

volatile, electrically erasable)

PIC24HJ32GP202 has 32Kbytes/11264

instructions, architecture supports

24Mbytes/4M instructions)On-chip Random Access Memory

(RAM) , volatile

PIC24HJ32GP202 has 2048 bytes,

architecture supports up 65536 bytes

Clock speed DC to 80 MHz

16-bit Architecture General purpose registers, 71 instructionsnot including addressing mode variants

On-chip modules Async serial IO, I2C, SPI, A/D, three 16-

bit timers, one 8-bit timer, comparator


4/75

4

PIC24 Core (Simplified Block Diagram)

17 x 17 Multipliernot shown

Program Counter

23address

Program Memory,non-volatile, up to4M words (4M x 24)

DOUT

24

ALU

DataMemInst. Reg

16

16

16

Data

16 address

16

16

16The inst ruct ion registercontains themachine code of the instruction currently

being executed.ALU (Arithmetic Logic Unit) is 16 bits wide,can accept as operands working registersor data memory.

16 x 16WorkingReg array


5/75

5

Memory Organization

Memory on the PIC24 C family is split into two types:Program Memoryand Data Memory.

PIC24 instructions are stored in program memory,which is non-volatile(contents are retained when poweris lost).

A PIC24 instruction is 24 bitswide (3 bytes).PIC24HJ32GP202 program memory supports 11264instructions; the PIC24 architecture can support up to

4Minstructions.PIC24 data is stored in data memoryand is amaximum size of 65536 x 8. Data memory is volatile(contents are lost when power is lost).


6/75

Program Memory

Locations 0x000000- 0x0001FF reserved, User program starts atlocation 0x000200.

Copyright Delmar Cengage Learning 2008. All Rights Reserved.

From: Reese/Bruce/Jones, Microcontrollers: From Assembly to C with the PIC24 Family.

Phantom byte: upper byte ofprogram memory that does notexist. Any read of the phantombyte return 0x00 and a write tothe phantom byte has no effect


7/75

7

Data Memory Organization


From: Reese/Bruce/Jones, Microcontrollers: From Assembly to C with the PIC24 Family .

Data memory forPIC24HJ32GP202


8/75

8

Special Function Registers (SFRs)

Special Function Registers (SFR) are

addressed like data memory locations.

SFRs live in the address range 0x0000 to 0x07FEin data memory. (check datasheet).

They are of two types: Core RegistersandPeripheralSFRs.

Peripheral SFRs are tied to processor

subsystems/peripherals (like the serial interface,or the analog-to-digital converter).

Peripheral SFR are used as control registersand

data registersfor processor subsystems.


9/75

9

Special Function Registers (SFRs)

Core Registers are used for storage of temporary

data; they are not used by the processorsubsystems.

These are sometimes referred to as GPRs

(general purpose registers). Sometimes alsoreferred to as file registers.

We typically refer to SFRs by name (W0, T3CON,

STATUS, etc) instead of by address.


10/75

8-bit, 16-bit, 32-bit Data

10

We will deal with data that is 8 bits, 16 bits(2 bytes), and 32 bits(4bytes) in size. Initially we will use only 8 bit and 16 bit examples.

Size Unsigned Range

8-bits 0 to 28-1 (0 to 255, 0 to 0xFF)

16-bit 0 to 216

-1 (0 to 65536, 0 to 0xFFFF)32-bit 0 to 232-1 (0 to 4,294,967,295), 0 to

0xFFFFFFFF)

The lower 8bits of a 16-bit value or of a 32-bit value is known as the Least

Significant Byte(LSB).

The upper 8bits of a 16-bit value or of a 32-bit value is known as the MostSignificant Byte(MSB).


11/75

Storing Multi-byte Values in Memory

11

16-bit and 32-bit values are stored in memory from least significant byte tomost significant byte, in increasing memory locations (little endianorder).


From: Reese/Bruce/Jones, Microcontrollers: From Assembly to C with the PIC24Family.


12/75

Data Transfer Instruction

Copiesdata from Source (src) location to Destination (dst)Location:

(src) dst ()read as contents of

This operation uses two operands.

The method by which an operandADDRESSis specified iscalled the addressingmode.

There are many different addressing modesfor the PIC24.

We will use a very limited number of addressing modes in ourinitial examples.

12


13/75

Data Transfer Instruction Summary

Key:MOV{.B} #lit8/16, Wnd PIC24 assemblylit Wnd Data transfer

Yellow shows varyingforms of the sameinstruction

f: near memory (08095) fALL: all of memory (065534) 13

8- or 16-bit literal: #lit8 or #lit16


14/75

V 0.2 14

MOV{.B } Wso, Wdo InstructionCopy contents of Wso register to Wdo register. General form:

mov{.b} Wso, Wdo (Wso) Wdo

Wso is one of the 16 working registersW0 through W15 (sindicatesWso is an operand sourceregister for the operation).

Wdo is one of the 16 working registers W0 through W15(dindicatesWdo is an operand destinationregister for the operation).

mov W3, W5 (W3) W5 (word operation)mov.b W3, W5 (W3.LSB) W5.LSB (byte operation)

Contents of working register W3copied toworking register W5.

This can either be a wordor byteoperation. The term copyis used

here instead of moveto emphasize that Wso is left unaffectedby theoperation.

The addressing mode used for both the source and destination operands iscalled register direct. The movinstruction supports other addressing modeswhich are not shown.

14


15/75

MOV Wso, Wdo Instruction Execution


From: Reese/Bruce/Jones, Microcontrollers: From Assembly to C with the PIC24 Family. 15


16/75

MOV Wso, Wdo Instruction Format



16


17/75

17

MOV Wns , fA ll Instruction

Copy contents of Wns register to data memory location f.General form:

MOV Wns, fAll (Wns) fAll

fis a memory locationin data memory, Wns is one of the 16 workingregisters W0 through W15 (sindicates Wns is an operand sourceregister for the operation)

MOV W3, 0x1000 (W3) 0x1000

Contents of register W3 copied to data memory location 0x1000.This instruction form onlysupports WORDoperations.

The source operand uses register directaddressing, while the destinationoperand uses file registeraddressing.

File registersis how Microchip refers to data memory.


18/75

18

MOV Wns , fAll Instruction Execution


From: Reese/Bruce/Jones, Microcontrollers: From Assembly to C with the PIC24Family.


19/75

19

MOV Wns , fAllInstruction Format




20/75

20

MOV fAll, Wnd Instruction

Copy contents of data memory location f to register Wnd. General form:

MOV fAll, Wnd (f) Wnd

fis a memory location in data memory, Wnd is one of the 16 workingregisters W0 through W15 (dindicates Wnd is an operand destinationregister for the operation).

MOV 0x1000, W3 (0x1000) W3

Contents of data memory location 0x1000 copied to W3.

() is read as Contents of.

This instruction form onlysupports a word operation.


21/75

21

MOV fA ll, Wnd Instruction Execution




22/75

A Note on Instruction Formats

The instruction formats (machine code) ofsome instructions will be presented forinformational purposes However, studying the machine code formatsof

the instructions is not a priority; understanding

instruction functionalitywill be emphasized. Allinstruction formatscan be found in the

dsPIC30F/dsPIC33F Programmers Referencemanual from Microchip

The PIC24 familyis a subset of thedsPIC30F/dsPIC33FF instruction setthe PIC24family does notimplement the DSPinstructions.

22


23/75

23

MOV{.B} WREG, f Instruction

Copy content of WREG (default working register) to data memorylocation f. General form:

MOV{.B} WREG, f (WREG) f

This instruction provides upward compatibilitywith earlierPIC C.WREG is register W0, and f is a location within the first 8192 bytes ofdata memory (neardata memory)

MOV WREG, 0x1000 (W0) 0x1000

Contentsof register W0 copied to data memory location0x1000.

Can be used for either WORD or BYTE operations:

MOV WREG, 0x1000 word operation

MOV.B WREG, 0x1001 lower 8-bits of W0 copied to 0x1001

Word copy must be to even (word-aligned) location.

Note: The previouslycovered MOV Wns, f instruction cannotbe used

for byteoperations!


24/75

24

MOV.B WREG, f Instruction Execution

A byte copy operation is shown.




25/75

25

MOV{.B} WREG, f Instruction Format


From: Reese/Bruce/Jones, Microcontrollers: From Assembly to C with the PIC24


26/75

26

MOV{.B } f {,WREG} Instruction

Copy contents of data memory location f to WREG (default workingregister) . General form:

MOV{.B} f, WREG (f) WREG

MOV{.B} f (f) f

This instruction provides upward compatibilitywith earlier PIC C.WREGis register W0, and f is a location within the first 8192 bytes ofdata memory (neardata memory)

Can be used for either WORD or BYTE operations:

MOV 0x1000, WREG word operation

MOV.B 0x1001, WREG only lower 8-bits of W0 are affected.MOV 0x1000 Copies contents of 0x1000 back to

itself, will see usefulness of this later

Wordcopy must be from even (word-aligned) data memory location.

Note: The MOV f,Wnd instruction cannotbe used for byteoperations!


27/75

MOV{.B } f {,WREG} Format

27Copyright Delmar Cengage Learning 2008. All Rights Reserved.From: Reese/Bruce/Jones, Microcontrollers: From Assembly to C with the PIC24 Family .

O { } f G I t ti


28/75

28

MOV.{B } f, WREG Instruction

Execution




29/75

29

Move a literal into a Working RegisterMoves a literalinto a working register. The #indicates the numeric valueis aliteral, and NOTa memory address.

General form:

MOV #lit16, Wnd lit16 Wnd (word operation)

MOV.B #lit8, Wnd lit8 Wnd.lsb (byte operation)

The source operand in these examples use the immediateaddressing mode.

Examples:

MOV #0x1000, W2 0x1000 W2

MOV.B #0xAB, W3 0xAB W3.lsb


30/75

More on Literals

Observe that the following twoinstructions are very different!

MOV #0x1000, W2 after execution, W2=0x1000

after execution, W2 = (0x1000), thecontents of memory location 0x1000

MOV 0x1000,W2

30


31/75

MOV Literal Execution



32/75

MOV Literal Instruction Formats


I di t Add i


33/75

Indirect Addressing

MOV with indirectAddressing:

mov{.b} [Wso], [Wdo] ((Wso)) (Wdo)

[](brackets) indicate indirectaddressing.SourceEffective Address(EAs) is the content of Wso, or (Wso).Destination Effective Address(EAd) is the content of Wdo, or (Wdo).

The MOV instruction copies the content of the Source Effective Address to theDestination Effect Address, or:

(EAs) EAd

which is:

((Wso)) (Wdo)

33


34/75



Indirect Addressing

MOV Example

34


35/75

Why Indirect Addressing?

35

The instruction:

mov [W0], [W1]

Allowsus to do a memory-memory copywith oneinstruction!

The following is illegal:

mov 0x1000, 0x1002

Instead, would have to do (2 instructions):

mov 0x1000, W0

mov W0, 0x1002


36/75

Indirect Addressing Coverage

There are six formsof indirect addressing

The need for indirect addressing makes the mostsense when covered in the context of Cpointers This is done in Chapter 5

At this time, you will only need to understand thesimplestform of indirect addressing, which isregister indirectas shown on the previous twoslides.

Mostinstructions that support register directfor anoperand, also support indirectaddressing as wellfor the same operand However, must check PIC24 datasheet and book to

confirm.


37/75


38/75


39/75

39

SUB{.B } Wb, Ws , Wd Instruction

Three operand subtraction, register-to-register form:

SUB{.B} Wb, Ws, Wd (Wb)(Ws) WdWb, Ws, Wd are any of the 16 working registers W0-W15.

Be careful:

while ADD Wx, Wy, Wz gives the same result as ADD Wy, Wx, Wz

The same is not true for

SUB Wx, Wy, Wz versus SUB Wy, Wx, Wz

SUB W0, W1, W2 (W0)(W1) W2

SUB W1,W0, W2 (W1)(W0) W2

SUB.B W0, W1, W2 Lower 8 bits of W0, W1 aresubtracted and placed in the lower 8-

bits of W2

SUB{ B} Wb Ws Wd Execution


40/75

SUB{.B } Wb, Ws, Wd Execution

40




41/75

Subtraction/Addition with Literals

Three operand addition/subtraction with literals:ADD{.B} Wb, #lit5, Wd (Wb) +#lit5 Wd

SUB{.B} Wb, #lit5, Wd (Wb)#lit5 Wd

#lit5 is a 5-bit unsigned literal; the range 0-31. Provides a convenientmethod of adding/subtracting a small constantusing a single instruction

Examples

ADD W0, #4, W2 (W0) + 4 W2

SUB.B W1,#8, W3 (W1)8 W3

ADD W0, #60, W1 illegal, 60 is greater than 31!


42/75

42

ADD{.B } f {,WREG} Instruction

Two operand addition form:

ADD{.B} f (f) + (WREG) fADD{.B} f, WREG (f) + (WREG) WREG

WREG is W0, f is limited to first 8192 bytes of memory.

One of the operands, either f or WREG is always destroyed!

ADD 0x1000 (0x1000) + (WREG) 0x1000

ADD 0x1000,WREG (0x1000) + (WREG) WREG

ADD.B 0x1001, WREG (0x1001) + (WREG.lsb) WREG.lsb


43/75

ADD{ B} f { WREG} Execution


44/75

44

ADD{.B } f {,WREG}Execution




45/75

Increment


46/75

V 0.2 46

IncrementIncrement operation, register-to-register form:

INC{.B} Ws, Wd (Ws) +1 Wd

Increment operation, memory to memory/WREG form:

INC{.B} f (f) + 1 f

INC{.B} f, WREG (f) + 1 WREG

(fmust be in first 8192 locations of data memory)

Examples:

INC W2, W4 (W2) + 1 W4

INC.B W3, W3 (W3.lsb) + 1 W3.lsb

INC 0x1000 (0x1000) +1 0x1000

INC.B 0x1001,WREG (0x1001)+1 WREG.lsb

Decrement


47/75

V 0.9

DecrementDecrement operation, register-to-register form:

DEC{.B} Ws, Wd (Ws)1 Wd

Increment operation, memory to memory/WREG form:

DEC{.B} f (f)1 f

DEC{.B} f, WREG (f)1 WREG

(fmust be in first 8192 locations of data memory)

Examples:

DEC W2, W4 (W2)1 W4

DEC.B W3, W3 (W3.lsb)1 W3.lsb

DEC 0x1000 (0x1000)1 0x1000

DEC.B 0x1001,WREG (0x1001)1 WREG.lsb

47


48/75

How is the instruction register loaded?

17 x 17 Multipliernot shown

Program Counter

23

addressProgram Memory,

non-volatile, up to

4M words (4M x 24)DOUT

24

ALU

DataMemInst. Reg

16

16

16

Data

16 address

16

16

16The Program coun tercontains theprogram memory address of the instruction

that will be loaded into the instructionregister . After reset, the first instructionfetched from location 0x000000inprogram memory, i.e., the program counteris reset to 0x000000.

16 x 16Working

Reg array

48


49/75

Program Memory Organization

An instruction is 24 bits (3 bytes). Program memory should be viewed as

words (16-bit addressable), with the upper byte of the upper word of aninstruction always reading as 0. Instructionsmust start on even-wordboundaries. Instructions are addressedby the Program counter(PC).

49Figure adapted with permission of the copyright owner,Microchip Technology, Incorporated. All rights reserved.

Goto location (goto )


50/75

50

Goto location (goto)How can the program counter be changed?

A GOTOinstruction is an unconditional jump.Copyright Delmar Cengage Learning 2008. All Rights Reserved.


Valid addressing modes


51/75

51

Valid addressing modes.What are valid addressing modes for instructions?

The definitive answer can be found in Table 19-2 of the PIC24H32GP202

datasheet.




52/75

What does Wso, Wsd, Wnetc. mean?

52

Table 19-1: Symbols used in opcode descriptions (partial list)

MOV Wso, Wdo


53/75

ADD forms

53

Legal:

ADD W0, W1, W2ADD W0, [W1], [W4]

Illegal:

ADD [W0],W1,W2 ;first operand illegal!

ADD Wb, Ws, Wd

Vid t t i l


54/75

54

Video tutorials

A number of videos illustrate important concepts; all are listed on the video pageat http://www.reesemicro.com/site/pic24micro/Home/pic24-video-tutorials-1.

Available tutorials, which cover topics on the following pages of these lecturenotes:

MPLAB IDE introductionathttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/mplab_assem/index.htm

A simple assembly language programathttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro/index.htm

Simulation of this programathttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro2/index.htm

Converting the program from 8 to 16 bitsathttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro3/index.htm

A Si l P
http://www.reesemicro.com/site/pic24micro/Home/pic24-video-tutorials-1http://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/mplab_assem/index.htmhttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro/index.htmhttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro2/index.htmhttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro3/index.htmhttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro3/index.htmhttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro2/index.htmhttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/assem_intro/index.htmhttp://www.ece.msstate.edu/courses/ece3724/main_pic24/videos/mplab_assem/index.htmhttp://www.reesemicro.com/site/pic24micro/Home/pic24-video-tutorials-1


55/75

55

A Simple Program

C Program equivalent

#define avalue 100uint8 i,j,k;

i = avalue; // i = 100

i = i + 1; // i++, i = 101

j = i; // j is 101

j = j - 1; // j--, j is 100

k = j + i; // k = 201

In this class, will present programs in Cform, then translate (compile) them toPIC24 C assembly language.

A uint8 variable is 8 bits(1 byte)


56/75

56

Where are variables stored?

When writing assembly language, can use any free data memory location tostore values, it your choice.

A logical place to begin storing data in the first free location in data memory,which is 0x0800 (Recall that 0x0000-0x07FF is reserved for SFRs).

Assign ito0x0800,jto 0x0801, and kto 0x0802. Other choices could be made.

C to PIC24 Assembly


57/75

57

Cto PIC24 Assembly

Comments: The assembly language program operation is not veryclear. Also, multiple assembly languagestatements are needed forone Clanguage statement. Assembly language is more pr imi t ive(operations less powerful) than C.




58/75

58

PIC24 Assembly to PIC24 Machine Code

Could performthis step manuallybydetermining the instruction formatfor eachinstruction from the data sheet.

Much easier to let a program called anassemblerdo this step automatically

The MPLABIntegrated Design Environment(IDE) is used to assemble PIC24 programsand simulate them Simulate means to execute the program without

actually loading it into a PIC24 microcontroller

.include "p24Hxxxx.inc"l b l t


59/75

59

.global __reset

.bss ;reserve space for variablesi: .space 1j: .space 1k: .space 1.text ;Start of Code section__reset: ; first instruction located at __reset label

mov #__SP_init, W15 ;;initialize stack pointermov #__SPLIM_init,W0mov W0,SPLIM ;;initialize Stack limit reg.avalue = 100

; i = 100;mov.b #avalue, W0 ; W0 = 100

mov.b WREG,i ; i = 100; i = i + 1;

inc.b i ; i = i + 1; j = i

mov.b i,WREG ; W0 = imov.b WREG,j ; j = W0

; j = j 1;

dec.b j ; j= j1

; k = j + imov.b i,WREG ; W0 = iadd.b j,WREG ; W0 = W0+j (WREG is W0)mov.b WREG,k ; k = W0

done:goto done ;loop forever

This file can beassembled by the

MPLABassemblerinto PIC24 machinecode and simulated.

Labels used formemory locations

0x0800 (i), 0x0801(j),0x0802(k) to increasecode clarity

mptst_byte.s

I l d fil th t d fi i


60/75

60

mptst_byte.s(cont.)

.include "p24Hxxxx.inc"

.global __reset

.bss ;reserve space for variablesi: .space 1

j: .space 1k: .space 1

Include filethat defines variouslabelsfor a particular processor.. includeis an assembler directive.

The .bssassembler directive

indicates the following shouldbe placedin data memory. Bydefault, variables are placedbeginning at the first freelocation, 0x800.

The .spaceassemblerdirectivereserves space inbytesfor the named variables.i, j, k are labels, and labels arecase-sensitive and must be

followed by a :(colon).

An assemb ler direct iveis not a PIC24instruction, but an instruction to theassembler program. Assembler

directives have a leading .period, andare not case sensitive.

Declare the __reset label asglobalit is is needed by linkerfor defining program start

t t b t ( t )


61/75

61

mptst_byte.s(cont.)

.text__reset: mov #__SP_init, W15

mov #__SPLIM_init,W0mov W0,SPLIM

avalue = 100

The equal sign is an assembler directivethatequates a labelto a value.

.textis an assembler directivethat says what followsis code.Our first instructionmust belabeled as __reset.

These move instructioninitializes the stack pointer and

stack limit registersthis will bediscussed in a later chapter.


62/75

62

mptst_byte.s(cont.)

; i = 100;mov.b #avalue, W0 ; W0 = 100mov.b WREG,i ; i = 100

; i = i + 1;

inc.b i ; i = i + 1; j = i

mov.b i,WREG ; W0 = imov.b WREG,j ; j = W0

; j = j 1;dec.b j ; j= j 1

; k = j + imov.b i,WREG ; W0 = iadd.b j,WREG ; W0 = W0+j (WREG is W0)

mov.b WREG,k ; k = W0

The use of labelsand commentsgreatly improves the clarityof theprogram.

It is hard to over-comment anassembly language program ifyou want to be able tounderstand it later.

Strive for at least a commentevery other line; refer to lines


63/75


64/75

Clock Cycles vs. Instruction Cycles


65/75

65

y yThe clock signal used by a PIC24 C to control instruction execution can begenerated by an off-chip oscillatoror crystal/capacitor network, or by usingthe internal RC oscillator within the PIC24 C.

For the PIC24H family, the maximum clock frequency is 80 MHz.

An instruction cycle (Fcy)is twoclock (Fosc)cycles.

A PIC24 instruction takes 1 or 2instruction (Fcy)cycles, depending on theinstruction (see Table 19-2, PIC24HJ32GP202 data sheet). If an instruction

causes the program counter to change (i.e, GOTO), that instruction takes 2instruction cycles.

An add instruction takes 1 instruction cycle. How much time is this if theclock frequency (Fosc) is 80 MHz ( 1 MHz = 1.0e6 = 1,000,000 Hz)?

1/frequency = period, 1/80 MHz = 12.5 ns (1 ns = 1.0e-9 s)1 Add instruction @ 80 MHz takes 2 clocks * 12.5 ns = 25 ns (or 0.025 us).

By comparison, an Intel Pentium add instruction @ 3 GHz takes 0.33 ns (330 ps).An Intel Pentium could emulate a PIC24HJ32GP202 faster than a PIC24HJ32GP202can execute! But you cant put a Pentium in a toaster, or buy one from Digi-key for

$5.00.

Important!!!!!!!

How long does mptst_byte.s take to execute?


66/75

g p _ y

66

InstructionCycles

mov #__SP_init, W15 1

mov #__SPLIM_init,W0 1

mov W0,SPLIM 1

mov.b #avalue, W0

1

mov.b WREG,i 1

inc.b i 1

mov.b i,WREG 1

mov.b WREG,j 1

dec.b j 1

mov.b i,WREG 1

add.b j,WREG 1

mov.b WREG,k 1

Total

12

Beginning at the __reset label, and ignoring the gotoat the end, takes 12instruction cycles, which is 24 clock cycles.

Wh t if d 16 bit i bl


67/75

What if we used 16-bit variables

instead of 8-bit variables?

67



i = avalue; // i = 2047

i = i + 1; // i++, i = 2048

j = i; // j is 2048

j = j - 1; // j--, j is 2047

k = j + i; // k = 4095

A uint16 variable is 16bits (1 byte)


68/75


69/75

An Alternate Solution (16-bit variables)

69



i = avalue; // i = 2047i = i + 1; // i++, i = 2048j = i; // j is 2048j = j - 1; // j--, j is 2047

k = j + i; // k = 4095

;Assign variables to registers;Move variables into registers.;use register-to-register operations forcomputations;

;write variables back to memory

;assign i to W1, j to W2, k to W3

mov #2047,W1 ; W1 (i) = 2047inc W1,W1 ; W1 (i) = W1 (i) + 1mov W1,W2 ; W2 (j) = W1 (i)dec W2,W2 ; W2 (j) = W2 (j) -1add W1,W2,W3 ; W3 (k) = W1 (i) + W2 (j);;write variables to memorymov W1,i ; 0x800 (i) = W1

mov W2,j ; 0x802 (j) = W2mov W3,k ; 0x804 (k) = W3

Previous approach took 9instructions, this one took 8instructions. In this case, this

approach is more efficient!

How long does mptst_word.s take to execute?


70/75

70

Instruction

Cycles

mov #__SP_init, W15 1

mov #__SPLIM_init,W0 1

mov W0,SPLIM 1

mov #avalue, W0

1

mov WREG,i 1

inc i 1

mov i,WREG 1

mov WREG,j 1

dec j 1

mov i,WREG 1

add j,WREG 1

mov WREG,k 1

Total

12

Ignoring the gotoat the end, takes 12 instruction cycles, which is 24 clockcycles.

16-bit operations versus 8-bit


71/75

16-bit operations versus 8-bit

71

The 16-bit versionof the mptstprogram requires the same number ofinstructionbytesand the samenumber of instruction cyclesas the 8-bitversion.

This is because the PIC24family is a 16-bitmicrocontroller; its naturaloperation size is 16 bits, so 16-bit operationsare handled as efficientlyas 8-bitsoperations.

On an 8-bit processor, like the PIC18family, the 16-bitversion wouldtake roughly doublethe number of instructions and clock cycles as the8-bitversion.

On the PIC24, a 32-bitversion of the mptstprogram will takeapproximately twicethe number of instructions and clock cycles as the16-bit version. We will look at 32-bit operations later in the semester.

Review: Units


72/75

72

Review: UnitsIn this class, units are always used for physical quantity:

Time Frequency

milliseconds (ms = 10-3s) kilohertz (kHz = 103Hz)

microseconds (

s = 10-6s) megahertz (MHz = 106Hz)

nanoseconds (ns = 10-9s) gigahertz (GHz = 109Hz)

When a time/frequency/voltage/current quantity is asked for, I will always ask forit in some units. Values for these quantities in datasheets are ALWAYS given inunits.

For a frequency of 1.25 kHz, what is the period in s?

period = 1/f = 1/(1.25 e3) = 8.0 e4 secondsUnit conversion= 8.0e-4 (s) * (1e6 s)/1.0 (s) = 8.0e2 s = 800 s

PIC24H Family


73/75

73

PIC24H Family

Microchip has an extensive line of PICmicro

microcontrollers, with the PIC24 familyintroduced in 2005.

The PIC16and PIC18are olderversions of thePICmicrofamily, have been several previous

generations. Do not assumethat because something is

done one way in the PIC24, that it is the mostefficient method for accomplishingthat action.

The datasheet for the PIC24HJ32GP202 isfound on the LMS.

Some PICMicros that we have used


74/75

V 0.2 74

Features 16F87x

(Fall 2003)

PIC18F242

(Summer 2004)

PIC24H

(Summer 2008)

Instructionwidth

14 bits 16 bits 24 bits

Program

memory

8K instr. 8K instructions ~10K instructions

DataMemory

368 bytes 1536 bytes 2048 bytes

Clock speed Max 20 MHz, 4

clks=1instr

Max 40 MHz

4 clks=1instr

Max 80 MHz

2 clks=1 instr

Architecture Accumulator, 8-

bit architecture

Accumulator, 8-bit

architecture General purposeregister, 16-bit

architecture

The PIC24H can execute about 6x faster than the PIC18F242 previouslyused in this class.

Wh t d d t k ?


75/75

What do you need to know?

Understand the PIC24 basic architecture(program and data memory organization)

Understand the operation of mov, add, sub, inc,dec, goto instructionsand their various

addressing modeforms Be able to convert simple C instruction

sequences to PIC24 assemblylanguage

Be able to assemble/simulate a PIC24 C assembly

language program in the MPLAB IDE

Understand the relationship between instructioncycles and machine cycles

Documents

EE353 L03 - PIC24 Family.ppt