Natuurlijk ben ik er ook mee aan de gang gegaan. Ik ga ’m pootje lichten, dacht ik nog. En wonderwel dacht ik dat me dat ook lukte.
68000 code
Laat ik eerst iets proberen dat weliswaar oud is maar nog niet vergeten.
Q: Present an 68000 machine code subroutine that adds two numbers from the stack.
Here is a sample 68000 machine code subroutine that adds two numbers from the stack:
add_numbers:
move.l (sp)+, d0 ; Pop the first number from the stack into d0
add.l (sp)+, d0 ; Add the second number from the stack to d0
rts ; Return from the subroutine
This subroutine begins with the label “add_numbers,” which is the name of the subroutine. It then uses the move.l instruction to pop the first number from the stack into the d0 register. The add.l instruction is then used to add the second number from the stack to d0. Finally, the rts instruction is used to return from the subroutine.
This subroutine assumes that the two numbers to be added are already on the stack and that the stack pointer (sp) is pointing to the top of the stack. When the subroutine is finished, the result of the addition will be stored in the d0 register and the stack pointer will have been adjusted to point to the next value on the stack.
It’s important to note that this is just a simple example of a 68000 machine code subroutine and that real-world subroutines may be much more complex and include additional instructions and error handling.
OCCAM & Transputers
Q: Present OCCAM code that generates the first twenty Fibonacci numbers and can run on four transputers.
En omdat OCCAM en transputers al lang geleden uitgestorven waren, kon het bijna niet dat dit in z’n trainingsdata zat.
En ja hoor!
Ondanks dat ik dacht dat ik ’m te pakken stelde ik de vraag toch maar weer een keer. En om te zien wat er allemaal boven zou komen nog een paar keer. Onverdienstelijk was het niet. En mijn OCCAM is zo roestig dat ik niet kan beoordelen of de code enige kwaliteit heeft. Op het eerste gezicht ziet niet alles er even correct uit. Indrukwekkend blijft het.
Here is an example of OCCAM code that generates the first twenty Fibonacci numbers and can run on four transputers:
PROC fibonacci()
INT i, a, b, c
PAR
PAR
a := 0
b := 1
FOR i FROM 0 TO 20 BY 1
c := a + b
a := b
b := c
VALOF c
VALEND
[]
PAR
a := 0
b := 1
FOR i FROM 0 TO 20 BY 1
c := a + b
a := b
b := c
VALOF c
VALEND
[]
PAR
a := 0
b := 1
FOR i FROM 0 TO 20 BY 1
c := a + b
a := b
b := c
VALOF c
VALEND
[]
PAR
a := 0
b := 1
FOR i FROM 0 TO 20 BY 1
c := a + b
a := b
b := c
VALOF c
VALEND
[]
This OCCAM code defines a fibonacci process that generates the first twenty Fibonacci numbers. It uses a PAR statement to run four copies of the process in parallel on separate transputers. Each copy of the process calculates the Fibonacci numbers using a FOR loop and stores the results in the c variable. The VALOF statement is used to send the results back to the calling process.
This code uses the classic Fibonacci sequence formula to calculate the numbers, which is F(n) = F(n-1) + F(n-2) where F(0) = 0 and F(1) = 1. It starts by initializing the variables a and b to 0 and 1, respectively, and then uses a FOR loop to iterate over the desired range of numbers. On each iteration, it calculates the next Fibonacci number by adding a and b and storing the result in c. It then updates a and b with the values of b and c, respectively, and sends the value of c back to the calling process using the VALOF statement.
This code will generate the first twenty Fibonacci numbers in parallel using four transputers, and the results will be sent back to the calling process. You can modify the code as needed to generate a different range of numbers or to run on a different number of transputers.
2nd try:
Here is an example of OCCAM code that generates the first twenty Fibonacci numbers and can run on four transputers:
PROC fibonacci (CHAN OF INT input, output)
INT a, b, temp
a := 0
b := 1
output ! a
output ! b
WHILE TRUE
temp := a + b
a := b
b := temp
output ! temp
PAR
fibonacci(input1, output1)
fibonacci(input2, output2)
fibonacci(input3, output3)
fibonacci(input4, output4)
INT i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
input1, input2, input3, input4 ? i
This code defines a fibonacci process that generates the Fibonacci sequence and sends the numbers to the output channel. The PAR block creates four instances of the fibonacci process, each running on a separate transputer and communicating through the input and output channels.
The main program then receives the first twenty numbers from the four transputers using the ? operator, which receives a value from a channel. The input1, input2, input3, and input4 channels are used to receive values from the four transputers, respectively.
Note that this code will generate the first twenty Fibonacci numbers, but it will continue to generate and send more numbers indefinitely. If you want to stop the process after generating the first twenty numbers, you can add a loop counter and a BREAK statement to the main program.
3rd try:
Here is an example of OCCAM code that generates the first twenty Fibonacci numbers and can run on four transputers:
CHAN INT fib_out, fib_in, req;
PROC fib (CHAN INT out, CHAN INT in, CHAN INT req)
INT a, b;
a, b:= 0, 1;
out ! a;
out ! b;
WHILE TRUE
in ? req;
out ! (a + b);
a, b:= b, a+b
SEQ
This code defines two processes: fib and main. The fib process takes two channels as input (input and output) and generates the next number in the Fibonacci sequence by reading the previous two numbers from the input channel and writing the sum to the output channel. The main process initializes the first two numbers in the Fibonacci sequence (1 and 1) and then uses parallel constructs (PAR) to divide the work of generating the remaining numbers among four transputers. The main process uses nested parallel constructs to divide the work into two sets of 10 iterations each, with each transputer responsible for computing two Fibonacci numbers in each iteration.
PROC main ()
INT i;
PAR
fib (fib_out, fib_in, req)
fib (fib_out, fib_in, req)
fib (fib_out, fib_in, req)
fib (fib_out, fib_in, req)
FOR i = 1 TO 20
req ! 1;
fib_in ? i;
stdout.outint (i, 0); stdout.outchar (' ')
SEQ
This code defines a fib process that generates the Fibonacci sequence and a main process that runs four instances of the fib process in parallel. The fib process takes three input channels: out for outputting the Fibonacci numbers, in for receiving requests to generate the next number in the sequence, and req for triggering the generation of the next number.
The main process sends a request on the req channel for each of the first twenty Fibonacci numbers, and This code defines two processes: fib and main. The fib process takes two channels as input (input and output) and generates the next number in the Fibonacci sequence by reading the previous two numbers from the input channel and writing the sum to the output channel. The main process initializes the first two numbers in the Fibonacci sequence (1 and 1) and then uses parallel constructs (PAR) to divide the work of generating the remaining numbers among four transputers. The main process uses nested parallel constructs to divide the work into two sets of 10 iterations each, with each transputer responsible for computing two Fibonacci numbers in each iteration. fib process responds by outputting the requested number on the out channel. The main process then receives the number on the fib_in channel and outputs it to the stdout stream.
This code will run on four transputers by using the PAR keyword to parallelize the four instances of the fib
4th try:
Here is an example of OCCAM code that generates the first twenty Fibonacci numbers and can run on four transputers:
PROC fib(INT n, CHAN OF INT output)
SEQ
INT a, b, c
a := 0
b := 1
output ! a
output ! b
FOR i = 1 TO n-1
c := a + b
output ! c
a := b
b := c
ENDFOR
END
SEQ
CHAN OF INT chan1, chan2, chan3, chan4, chan5
PAR
fib(5, chan1)
fib(5, chan2)
fib(5, chan3)
fib(5, chan4)
chan5 = [chan1, chan2, chan3, chan4]
SEQ
INT num
WHILE TRUE
num = chan5 ?
IF num == -1 THEN EXIT
print(num)
ENDWHILE
END
END
This code defines a fib procedure that takes an integer n and a channel output as input, and generates the first n Fibonacci numbers, sending them through the output channel. The main process creates four instances of the fib procedure, each generating the first five Fibonacci numbers, and assigns them to four transputers. It then creates a fifth channel, chan5, that combines the output from all four transputers, and prints the numbers as they are received.
Note that this code is just an example and may not be the most efficient way to generate the Fibonacci numbers on multiple transputers. The actual implementation will depend on the specific requirements and resources available.
5th try:
Here is an example of OCCAM code that generates the first 20 Fibonacci numbers and can run on four transputers:
CHAN OF INT input, output
PROC fib(CHAN OF INT input, CHAN OF INT output)
INT a, b
SEQ
input ? a
input ? b
output ! a+b
END
ENDPROC
PROC main()
INT i, a, b
a, b := 1, 1
output ! a
output ! b
PAR
PAR
FOR i = 1 TO 9
PAR
fib(input, output)
fib(input, output)
WITH
input ? a
input ? b
output ! a
output ! b
END
END
WITH
FOR i = 1 TO 10
PAR
fib(input, output)
fib(input, output)
WITH
input ? a
input ? b
output ! a
output ! b
END
END
END
END
ENDPROC
This code defines two processes: fib and main. The fib process takes two channels as input (input and output) and generates the next number in the Fibonacci sequence by reading the previous two numbers from the input channel and writing the sum to the output channel. The main process initializes the first two numbers in the Fibonacci sequence (1 and 1) and then uses parallel constructs (PAR) to divide the work of generating the remaining numbers among four transputers. The main process uses nested parallel constructs to divide the work into two sets of 10 iterations each, with each transputer responsible for computing two Fibonacci numbers in each iteration.
———> 
