TrueCrypt 7.0a's Performance, Analyzed

tomshardware.com

14 pointsdotpot4 comments

I really, really wish more reviews would use real-world scenarios (file copying/moving, boot-up time, application load times, game level load times, etc) when benchmarking hard drives instead of almost useless theoretical tests or benchmark suites that give you a large number at the end that means nothing. All this review tells me is that hardware decryption is faster than software decryption.

Yeah, would be great to get that statistics/benchmark as well, because now that does not making huge clue.

Question: would adding another hardware RNG to feed the random queue significantly boost performance here?

No.

http://www.truecrypt.org/docs/random-number-generator

The RNG pool is used for creating: the master key (for encrypting the data in the volume), temporary secondary key (for generating a pseudorandom stream to initially overwrite a volume's empty space), salt (for passphrase hashing, using pbkdf2), and (optionally) a keyfile.

That is all limited to the process of setting up a volume, and so the RNG pool is not used for typical operations on a volume after that volume is created.

Reply on news.ycombinator.com