I have migrated my project from Vivado 2015.4 to Vivado 2016.2. In Vivado 2016.2, after the output targets are generated, any action involving the Vivado IDE (GUI) becomes extremely slow.
For example, it takes several seconds just to mark one menu, or to open one window. It also seems to be worse with larger projects or when running over a network.
This slowness is indirectly caused by the number of runs in a project. Vivado is constantly polling each run to check status and message information in the run. As the number of runs increases, the time to check all of the runs becomes greater that the interval between checks.
This issue occurs most often with projects containing a Block Diagram (BD), where the output products for the BD are generated with the "OOC per IP" option. This creates a new run for each IP instance in the BD.
If the large number of project runs is primarily due to the BD output targets, and the design does not require that OOC be run on a per IP instance basis, the issue can be worked around by generating the output targets as Global or as one OOC per BD.
The attached patch (AR67729_vivado_2016_2_preliminary_rev1.zip) can be downloaded to correct this issue. With the patch, Vivado will dynamically adjust the frequency of the runs check to avoid the issue.
|Name||File Size||File Type|