高扇出寄存器没有被复制？

创建一个包含第二轮phys_opt_design命令的.tcl文件。
然后在Implementation Options> phys_opt_design tcl.post选项中指向此.tcl文件。
--------------------------------------------------
-----------------------不要忘记回答，kudo，并接受为解决方案.-------------
--------------------------------------------------
----------
在原帖中查看解决方案



以下为原文

Create a .tcl file that contains the second round of phys_opt_design command. Then point to this .tcl file in Implementation Options > phys_opt_design tcl.post option.
-------------------------------------------------------------------------
Don't forget to reply, kudo, and accept as solution.
-------------------------------------------------------------------------View solution in original post

嗨@tessitd
你尝试过max_fanout约束吗？
有些事情可能会使max_fanout无法正常工作，详细信息请参见https://www.xilinx.com/support/answers/62162.html
谢谢，维杰-----------------------------------------------
---------------------------------------------请将帖子标记为
一个答案“接受为解决方案”，以防它有助于解决您的查询。如果一个帖子引导到解决方案，请给予赞誉。



以下为原文

Hi @tessitd
Did you try max_fanout constraint?
There are few things that may get the max_fanout to not work as expected and details are in https://www.xilinx.com/support/answers/62162.html
 
Thanks,Vijay
--------------------------------------------------------------------------------------------
Please mark the post as an answer "Accept as solution" in case it helped resolve your query.
Give kudos in case a post in case it guided to the solution.

高扇出优化的工作原理如下：
1.高扇出网，在WNS的百分比内具有负松弛，被认为是复制的。
根据邻近度对负载进行群集，并为每个负载群集复制和放置驱动程序。
并非所有高扇出网都是复制的候选者。
或者，您可以在尝试再次路由设计之前添加phys_opt_design迭代以强制在特定网络上进行复制。
例如：phys_opt_design -force_replication_on_nets [get_nets [list netA netB netC]]
--------------------------------------------------
-----------------------不要忘记回答，kudo，并接受为解决方案.-------------
--------------------------------------------------
----------



以下为原文

High-Fanout Optimization works as follows:
1. High fanout nets, with negative slack within a percentage of the WNS, are considered for replication.
2. Loads are clustered based on proximity, and drivers are replicated and placed for each load cluster.
 
Not all high fanout nets are candidates for replication.
Alternatively, you can add an iteration of phys_opt_design to force replication on specific nets before trying to route the design again. For example:
phys_opt_design -force_replication_on_nets [get_nets [list netA netB netC]]
-------------------------------------------------------------------------
Don't forget to reply, kudo, and accept as solution.
-------------------------------------------------------------------------

维杰，
所以我会在Synthesis或P＆amp; R中设置这个约束？
那是为了max_fanout？
我可以轻松地将其设置为代码中的VHDL属性，因为我知道哪些信号先前负载很重。
我也很可能会对它们进行保留。
这是推荐的流量吗？
TomT ...



以下为原文

Vijay,
 
So I would set this constraint in Synthesis or in P&R?  That is for the max_fanout?  I could easily set this as a VHDL Attribute in the code as I know which signals are heavily loaded prior.  I most likely would put a KEEP on them as well.
 
Is this the recommended flow?
TomT...

美惠，
我早期注意到的一件事是我不能在-force_replication_on_nets命令中使用策略。
所以这似乎是一个多次通过流程来使这项工作。
现在当前的流程是混合的，我基本上每次都从头开始构建一个项目并遵循“项目”流程。
但实际上它都是通过TCL设置的。
我不是在做“非项目”流程。
所以我看到让这个版本的phys_opt_design流程起作用是有问题的。
当我对许多网络使用“AgressiveExplore”时，phys_opt_design正在复制，当我仍然有-0.500ns的松弛时，它只是没有修复这个区域，我可以直接归因于高扇出网消耗大约7.5ns的延迟。
所以我想保留“AgressiveExplore”策略，然后再添加第二遍phys_opt_design -force_replication。
我如何在基于PROJECT的流程中执行此操作？
谢谢，
TomT ...



以下为原文

Graces,
 
One thing I noticed early on is that I cannot use a Strategy with the -force_replication_on_nets command.  So this seems to be a multi-pass flow to make this work.
 
Now current flow is a hybrid, I basically build a project each time from scratch and follow the "Project" flow.  But in reality it is all setup via TCL.  I am NOT doing the "non-project" flow.  So I see getting this version of the phys_opt_design flow to work is problematic.
 
phys_opt_design is replicating when I use the "AgressiveExplore" for many nets, it just isn't fixing this one area when I still have a -0.500ns slack which I can directly attribute to the high fanout net consuming about 7.5ns of the delay.  So I would like to keep the Strategy of "AgressiveExplore" but then add a second pass through the phys_opt_design -force_replication.  How do I do that in a PROJECT based flow?
 
Thanks,
TomT...

创建一个包含第二轮phys_opt_design命令的.tcl文件。
然后在Implementation Options> phys_opt_design tcl.post选项中指向此.tcl文件。
--------------------------------------------------
-----------------------不要忘记回答，kudo，并接受为解决方案.-------------
--------------------------------------------------
----------



以下为原文

Create a .tcl file that contains the second round of phys_opt_design command. Then point to this .tcl file in Implementation Options > phys_opt_design tcl.post option.
-------------------------------------------------------------------------
Don't forget to reply, kudo, and accept as solution.
-------------------------------------------------------------------------