When inferring primitives or logic, the device utilization might be much larger than expectedin some cases.The design works fine insimulation, but XST appears to freeze when implementing, timing might be difficult to meet, or device utilization can be extremely high.
How can I resolve it?
This issue is primarily caused by not following the inference templates described in the XST User Guide.While a design might work properly in simulation, it does not necessarily mean that the HDL description is reasonable for the device architecture.The XST User Guide provides inference templates that should be followed when primitives are not being explicitly instantiated. XST is looking for specific template matches, so even if the HDL description is only slightly different, it might cause synthesis failures as previously described.
The XST User Guide can be found at: