是否可以使用AXI4流以某种方式从收发器中提取输入数据

长话短，未经编码的NRZ数据以5.6 GB / s的速度接收
简而言之，这是不可能做到的。
在5.6gbps时，您无法使用FPGA的“常规”（并行）I / O.
高速I / O（GTX / GTH）使用时钟数据恢复（CDR），因此必须使用协议以确保CDR保持锁定。
如上所述，你无法做你想做的事。
但是，你为什么要这样做？
这个5.6Gbps数据流来自哪里？
它代表什么？
我上面描述的问题非常普遍 - 我没有能够捕获此速率的未编码串行流的设备，因此我知道没有设备生成一个（因为生成没有设备可以接收的输出是没有意义的）
）。
我们需要绘制数据
这意味着数据在某种程度上可以被分组为单词（因为“绘制”一个比特流并没有意义。如果这是一堆单词，必须有一些协议允许你构建
串行数据提取单词...
因此，除非这是一个非常奇怪的（和实验性的）输出设备，否则它不太可能以此速率生成未编码的位串行流。
Avrum



以下为原文

Long story short, unencoded NRZ data is being received at 5.6 GB/s
 
In a nutshell, this can't be done.
 
At 5.6gbps you cannot use the "conventional" (parallel) I/O of the FPGA.
 
The high speed I/O (GTX/GTH) use clock data recovery (CDR), and hence must use a protocol in order to ensure that the CDR can remain locked.
 
So as described, you cannot do what you are trying to do.
 
But, why do you want to do this? Where is this 5.6Gbps datastream coming from? What does it represent? The problems I describe above are pretty much universal - I no of no device that can capture an unencoded serial stream of this rate, and hence I know of no device that generates one (since it is pointless to generate an output that no device can receive).
 
and we need plot the data
 
This implies that the data is somehow able to be grouped into words (because it doesn't really make sense to "plot" a stream of bits. If this is a bunch of words, there must be some protocol to allow you to frame the serial data to extract the words...
 
So, unless this is some very odd (and experimental) output device, it is unlikely that it really generates an unencoded bit serial stream at this rate.
 
Avrum