Makes no difference.
http://www.schneier.com/blog/archive...tack_on_a.html
We also expect that a careful analysis may reduce the complexities. As a preliminary result, we think that the complexity of the attack on AES-256 can be lowered from 2^119 to about 2^110.5 data and time.
Let's focus on time. Let's lower that to 2^110 for simplicity. Let's say with the plaintext the NSA can attempt one decryption operation in one nanosecond (one billionth of a second). Then to find the key takes:
2^110 / ( 10^9 * 3600 * 24 * 365.25 *10^15) = 41 quadrillion years.
Now let's say the NSA has spent $10 trillion to buy enough computers to build a massively parallel key cracker. Let's say each cracker costs just $10. So the NSA has one trillion computers to crack keys. So instead of
41 quadrillion years, the NSA takes a mere 41 thousand years.
And I've been extremely generous in my over estimation of the resources the NSA has. I suspect that it takes at least a microsecond, not a nanosecond, to try to decrypt some ciphertext and compare to the known plaintext. I suspect the NSA has spend closer to $100B on its key cracking hardware. And I suspect a key cracker node is closer to $100. So the number is likely closer to 41,000 * 1000 * 100 * 10, or 41 billion years to crack a key.
As I said, good luck with that.