Is Argon2 based on BLAKE2b or is it only using the BLAKE2b at the end?

adler

8/26/24, 11:45 PM

What's the exact meaning of 'Argon2 uses BLAKE2b'? Does it based on BLAKE2b? like chacha20 based on salsa20 kind of meaning or does Argon2 use the BLAKE2b with 12 rounds hash function at the end to generate a hash value? thank you.

131

1 + 0

blake2

argon2

Score:3

Crypto

bk2204

8/27/24, 1:49 AM

Argon2 uses BLAKE2b in several different ways, which can be seen in the Wikipedia article on it. It's first used to hash the parameters, including the password, and then used in a construct that expands an input to an effectively arbitrary length digest.

This arbitrary-length digest is used to expand the first two columns, and then to generate the final hash value. Assuming the final output is 64 bytes (512 bits) or fewer in length, then yes, it's effectively just BLAKE2b with the digest length (32-bit little endian) prepended to the message, possibly truncated.

There are other components, such as the G function, which is used for expanding the columns, which are based on BLAKE2b, but have some differences (notably the use of multiplications).

+ 3

adler

8/27/24, 8:49 AM

thanks, so it's not chacha based on salsa relationship, right? but at some point(s) in your first example Argon2 calls the 12 rounds BLAKE2b to hash the parameters, correct?

bk2204

8/27/24, 2:18 PM

Yes, there are actual invocations of BLAKE2b proper without any modifications. There are also, as I mentioned, some components that resemble BLAKE2b but are slightly different much as ChaCha20 resembles Salsa20. So we have both.

forest

8/27/24, 10:51 PM

The BLAKE2-based algorithm in the G function is called BlaMka (which is also used in an algorithm called Lyra2). It is just used for fast diffusion and isn't a secure hash on its own.

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Is Argon2 based on BLAKE2b or is it only using the BLAKE2b at the end?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.