LDPC Decoding: VLSI Architectures and Implementations
LDPC Decoding: VLSI Architectures and Implementations
LDPC Decoding: VLSI Architectures and Implementations
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Layered Decoder Architecture<br />
Our work proposed this for H matrices with regular<br />
mother matrices.<br />
Compared to other work, this work has several advantages<br />
1) No need of separate memory for P.<br />
2) Only one shifter instead of 2 shifters<br />
3) Value-reuse is effectively used for both Rnew <strong>and</strong> Rold<br />
4) Low complexity data path design-with no redundant data<br />
Path operations.<br />
5) Low complexity CNU design.<br />
r<br />
R<br />
( i)<br />
l,<br />
n<br />
=<br />
⎛<br />
f ⎜<br />
⎜<br />
⎝<br />
r<br />
[<br />
( )<br />
] ( ′<br />
i<br />
S l,<br />
n<br />
Q<br />
) ⎟ ⎟ ⎞<br />
l,<br />
n′<br />
,<br />
∀ n′<br />
= 1,2, L,<br />
k⎠<br />
r<br />
[<br />
( )] [ ]<br />
i<br />
S ( l,<br />
n)<br />
S ( l,<br />
n)<br />
( i−1)<br />
Q = P − R<br />
l,<br />
n<br />
r<br />
n<br />
r<br />
l,<br />
n<br />
r<br />
r ( r<br />
P<br />
n<br />
l,<br />
n<br />
l,<br />
n<br />
[ ]<br />
S l,<br />
n)<br />
[<br />
( )] i<br />
S ( l,<br />
n)<br />
( i)<br />
= Q + R<br />
Q=P-Rold<br />
Flash Memory Summit 2013<br />
Santa Clara, CA 20