If you see in other way around... bottom up approach for a chip means top down for that particular Block/core/submodule right. Thus need to close timing exclusively. Then try and extarct the timing models of the blocks and then perform full chip timing. This will save the run time also.