# Fill an array with assembler

I would like to fill an dynamic array of
arbitrary size with a value:

procedure FillWithValue(arr:pointer;n:longint;fval:double);assembler;
begin
asm
// ???
end;
end;

type tArr = array[0..0] of double;
pArr = ^tArr;

procedure Process;
var i:longint;
arr:pArr;
count:longint;
begin
count:=3000; // arbirary value
getmem(arr,count*sizeof(double));
for i:=0 to count-1
do begin
arr^[i]:=2*i;  // arbitrary start value
end;
i:=200;
ShowMessage(floattostr(arr^[i]));
// now fill with the whole array with 0.0
FillWithValue(arr,count,0.0);
freemem(arr);
end;

Can someone help me with the assembler code
to put fval into all elements of arr?

Thanks for help,

Gamba

###### Who is Participating?

Commented:
Force accepted

** Mindphaser - Community Support Moderator **
0

Commented:
why don't you simply use FillChar
I'm sure it's highly optimized :)

you could also write some code with FillChar and check the asm procedure it creates :)
0

Commented:
I think it's becoz he's trying to fill the array with double and not char :)
0

Commented:
umm yes ...
well filling the array with all 0.0 is the same as filling it with 0 chars :)
0

Commented:
Umm... true... I didn't see that part of filling with 0.0... lol

So Gamba, just use

FillChar(YourArray, 0, SizeOf(YourArray));
0

Commented:
PS: If it works, please give the points to Lee and not to me :)
0

Commented:
hehe tnx :)

anyway it should be like :
FillChar(YourArray, Length(YourArray) * SizeOf(Double), 0)
0

Commented:
Oops... here I go again... must be getting old ;)
0

Commented:

hi gamba,
why do you need assembler code? It is difficult to write assembler code for short operations that works faster then Delphi build. Delphi does excellent optimization and in some cases writing fastest code inside one procedure in assembler will slow down whole process. It is due to Delphi can't optimize procedure that calls procedure writen in assembler.

Just use this way to work with array of doubles:

var
A: Array of Double;
I: Integer;
begin

SetLength(A, 12);
for I := Low(A) to High(A) do
A[I] := I * 2;
end;

-------
Igor.
0

Author Commented:

You're right in that FillChar would be an alternative
for filling with zeros. But this was meant only as an
example for accessing the array via assembler.
What I want to do in the end is processing a huge
array of floating point values with assembler, thereby
accessing the FPU direct to do mathematical operations
on the array.
So I need the assembler code for getting started at all.

Gamba

0

Commented:
No comment has been added lately, so it's time to clean up this TA.
I will leave a recommendation in the Cleanup topic area that this question is:

PAQ and No Refund

Paul (pnh73)
EE Cleanup Volunteer
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.