08-05-2019 04:11 AM
Hello, I'm using the aurora 64/66b IP (Kintex 7, RX simplex @ 1.28Gbps) and I would like to modify it, to remove the GTX transceivers instantiation from the ip and keep only the actual aurora core.
The problem is, that on top of the aurora encoding we will have additional (custom) encoding thus, we will have to connect another block on the receiving side to perform the first decoding (and this block will now handle the GTX pins) and then feed the aurora block.
My question is, is this something that can be done or the handlng of the GTX transceivers is too deeply integrated in the ip and it would be easier to develop a new aurora decoder block?
If this is something feasible, could you give some advice on the implementation?
08-05-2019 11:33 PM
I'm afraid there is no such an option to seperate the GT and Aurora IP.
This is not supported.
If you really have to do that, you can put the First decoding in the *_multi_gt_i module, read the RXDATA_OUT, decode it and send the decoded rxdata out through the GT*_RXDATA_OUT.
Below is the snapshot.
this module is read only in vivado but you can edit it manually by a text editor.
08-06-2019 03:54 AM
Thank you for your reply, but unfortunatelly I don't think that this is an option for us.
We cannot modify the other block, it has to control the GTs and the aurora block has to follow after that.
So, we have to choose either to dig deep in the aurora ip and manually modify it to separate GT (but this would imply a really good understanding of everything that is happening inside the code) or develop a custom aurora block.