Hi,
Do you have something
more about synchronization between different clock domains?
Best Regards,
XiaYun
-----邮件原件-----
发件人:
owner-ethmac@opencores.org [mailto:owner-ethmac@opencores.org] 代表 Illan Glasner
发送时间:
2002年11月1日 4点37
收件人:
'ethmac@opencores.org'
主题: RE: [ethmac]
gray code
Gray format mean
that there is only one bit change at a time this mean that you can do a
comparision without the need to do complex synconization of moving the pointer
from one domain to another.
you can do for example something like :
wadd_gray_s <= #1 wadd_gray;
always @(posedge rclk or negedge rstn)
if (~rstn)
ordy <= #1
~((wadd_gray_s==radd_gray) || (re &&
(wadd_gray_s==radd_gray_next)) );
where wadd_gray come from wclk domain.
as compare to much more complicated such as store
and hold you will need to do in binary.
but even when using Gray it is usualy used only
for the flags while there is another set of counter used as the index for the
read/write and those are binary, therefore also you will see many time gray is
used only for small fifo's as the "overhead" logic for "full
syncroniztion" become smaller when you need to have large counter for
binary as well as for gray.
Another "advantage" for gray as compare
to the full syncornization is that syncronization methode always use pesemstic
approch which mean let say you have fifo with 8 word it might show full even
few clocks after the full level is no more full or it might show empty for few
clocks before it show not empty. however in Gray the delay is not exsisting or
can be 1 clock delay.
-----Original
Message-----
From: Madhu rao
[mailto:madhu_ether@yahoo.com]
Sent: Tuesday, October 29, 2002
11:13 PM
To: ethmac@opencores.org
Subject: [ethmac] gray code
Hai,
why
people prefer to store data in fifo as gray format instead of binary format. Is
it for speed improvent or any specific reasons.
Thanks in
advance.
Regards
Madhu
Do you
Yahoo!?
HotJobs
- Search new jobs daily now