Hi Nico,
Thanks for the reply, I tried C version, it is too slow. I also tried
to put an assembly block inside the double loops (written in C),
definitely faster than C, but not enough. That's why I want to use
assembly. Do you know what's wrong with my code? Thanks!
haodong
N. Coesel wrote:
At 18:18 12-10-05 -0400, you wrote:
Hi,
I have a question regarding msp430 assembly language, I am trying to do a
big integer multiplier, wrote the following code (only the first half
loop). The program is to compute c[] = a[] * b[]. When I compiles the
code, the compiler reports "inconsistent operand constraints in an `asm'",
I guess there is something wrong at indexed addressing, how can I fix the
code?
Just another approach to your problem:
Can you try writing the code in C and see if it is fast enough for your
needs? I think the code produced by the compiler is probably better than
your assembly code.
See the following example:
void test()
{
int i,j,a,b,c;
a=1;
b=2;
for(i=0; i<1000; i++) {
1f6: 0d 43 clr r13 ;
for(j=0;j<1000; j++) {
1f8: 3e 40 e7 03 mov #999, r14 ;#0x03e7
c=a*b;
1fc: 02 12 push r2 ;
1fe: 32 c2 dint
200: 03 43 nop
202: 92 43 32 01 mov #1, &0x0132 ;r3 As==01
206: a2 43 38 01 mov #2, &0x0138 ;r3 As==10
20a: 1f 42 3a 01 mov &0x013a,r15 ;0x013a
20e: 32 41 pop r2 ;
210: 3e 53 add #-1, r14 ;r3 As==11
212: 3e b0 00 80 bit #-32768,r14 ;#0x8000
216: f2 37 jge $-26 ;abs 0x1fc
218: 1d 53 inc r13 ;
21a: 3d 90 e8 03 cmp #1000, r13 ;#0x03e8
21e: ec 3b jl $-38 ;abs 0x1f8
}
}
}
220: 30 41 ret
Nico Coesel
-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Mspgcc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mspgcc-users
