Use computeProductBlockingSizes to compute blocking for both ShardByCol and ShardByRow cases.
1 file changed