I don't think we expect the blk* files to be identical between different clients. They can contain orphaned blocks for example. But the total size of the blk files should be at least as big as the bootstrap.dat file since the bootstrap.dat file doesn't contain orphaned blocks.
This makes me think you're doing more than simply "cat blk*.dat > bootstrap.dat", so how are you generating the bootstrap?
A bootstrap.dat file has to contain the blocks in the correct order, or it won't import. The blk*.dat files contain the blocks in the order they were received from peers, and so are often not in the correct order.
I create my bootstrap.dat files using the "dumpbootstrap" RPC in the client, which dumps the blocks out to a file in the correct order.