Retrieving 32bit Multiply Result

[skimask]

Is it possible to have a big_number = 1,000,000 – 2,000,000?
Nick

It's directly possible if you have the PBP 2.50a upgrade, which supports signed/32 bit LONG variables...

[mister_e]

for PIC18 only !

[skimask]

for PIC18 only !

Got my back everywhere, dontcha?

[Nicmus]

Thank you gents for your replies.
Unfortunately for me I’m dealing with a old application which employs a PIC16F876A and needs an upgrade to be able to handle a larger number of records.
The code space is not a problem so I guess I will have to figure a way to get around this problem.
Basically I receive a 8 bytes decimal number which is the total number of records. What I’m planning to do is create two loops.
First I will separate the incoming Big_number in Big_number_low (last four bytes (max 9999)) and Big_number_high (first four MS bytes (max 9999)).
The code in my last post should change to:

For index_1=0 to Big_number_low
HSEOUT [sentence(index_1)]
Next
For index_2=0 to Big_number_high
For index_3=0 to 9999
HSEROUT [sentence(index_3)]
Next
Next

I’m wandering if this will do or I need to look deeper into the problem.
Any comments will be appreciated.

Regards,

Nick

[Darrel Taylor]

Not sure how you're going to get a 1,000,000 element array with a PIC ...
Even with PBP 2.50

For index =0 to big_number
    Hserout [record(index),13,10]
    Pause 100
Next

Not enough RAM.
<br>

[Nicmus]

Thank you gents for your replies.
I’m sorry for not making it clear. I’m receiving each sentence one at the time and send them out. My problem is that I know how many I will receive (Big_number) and I need to keep track on how many I send out to match the Big_number. So it gets down to only handling a 6digi number.
Unfortunately for me I’m dealing with a old application which employs a PIC16F876A and needs an upgrade to be able to handle a larger number of records.
The code space is not a problem so I guess I will have to figure a way to get around this problem.
Basically I receive a 8 bytes decimal number which is the total number of records. What I’m planning to do is create two loops.
First I will separate the incoming Big_number in Big_number_low (last four bytes (max 9999)) and Big_number_high (first four MS bytes (max 9999)).
The code in my last post should change to:

For index_1=0 to Big_number_low
HSEOUT [sentence(index_1)]
Next
For index_2=0 to Big_number_high
For index_3=0 to 9999
HSEROUT [sentence(index_3)]
Next
Next

I’m wandering if this will do or I need to look deeper into the problem.
Any comments will be appreciated.

Regards,


Nick

[Darrel Taylor]

How are you going to get a 1,000,000 element array with a PIC?

For index_1=0 to Big_number_low
    HSEROUT [sentence(index_1)]
Next

Not enough RAM.
<br>

[Nicmus]

Hi Darrel,
I’m sorry for not presenting the complete picture.
I’m reading a .txt file from a USB media. The first line of this file gives me the total number of records in the file.
I’m reading the records one by one in a big loop:

Read Big_number
For index = 0 to Big_number
Read record
HSEROUT [record}
If index => Big_number then quit
Next

It all works fine if Big_number is limited to word size.
Now the requirement changed for larger number of records so I thought that your trick might apply here.
At the time I did that because I did not know, and I still don’t, to detect EOF.
Trying to read the device over the EOF will freeze the system.

Thank you for your interest.

Nick

[Darrel Taylor]

That makes things a bit easier ...

big_number  VAR WORD[2]   ; 32-bit variable 


ASM  
;---[load a 32-bit constant into a 32-bit variable]-----------------
MOVE?CN  macro Cin, Nout
    MOVE?CW  Cin & 0xFFFF, Nout     ; Low Word
    MOVE?CW  (Cin >> 16), Nout + 2  ; High Word
  endm
ENDASM


;-------------------------------------------------------------------
@  MOVE?CN  1000000, big_number

WHILE (big_number[1] > 0) OR (big_number[0] > 0)
    Read record
    HSEROUT [record]
    
    big_number[0] = big_number[0] - 1
    IF big_number[0] = $ffff THEN big_number[1] = big_number[1] - 1
WEND