PierMesh
/
direwolf
mirror of https://github.com/wb2osz/direwolf.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Small improvement of FX.25 RS encoder speed 

The old design walked the entire feedback array XORing all values, then
shifted them all down by one index. This design does the shift-by-one
while doing the XOR work to save a step.

Signed-off-by: Justin Brzozoski <justin.brzozoski@gmail.com>
This commit is contained in:
Justin Brzozoski 2020-06-09 22:51:59 -04:00
parent 5fd8120de7
commit e3f834f261
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/fx25_encode.c
View File

@ -69,15 +69,15 @@ void ENCODE_RS(struct rs * restrict rs, DTYPE * restrict data, DTYPE * restrict
for(i=0;i<NN-NROOTS;i++){
feedback = INDEX_OF[data[i] ^ bb[0]];
if(feedback != A0){ /* feedback term is non-zero */
/* Shift and XOR */
for(j=1;j<NROOTS;j++)
bb[j] ^= ALPHA_TO[MODNN(feedback + GENPOLY[NROOTS-j])];
}
/* Shift */
memmove(&bb[0],&bb[1],sizeof(DTYPE)*(NROOTS-1));
if(feedback != A0)
bb[j-1] = bb[j] ^ ALPHA_TO[MODNN(feedback + GENPOLY[NROOTS-j])];
bb[NROOTS-1] = ALPHA_TO[MODNN(feedback + GENPOLY[0])];
else
} else{
/* Shift only */
memmove(&bb[0],&bb[1],sizeof(DTYPE)*(NROOTS-1));
bb[NROOTS-1] = 0;
}
}
}