The current default cryptographic hash functions are serial, in that you can’t efficiently use a cluster computers to hash a 10TB file, say.
When do people think this change? Concretely, when will the hash function people default to (e.g., the one used to hash Linux kernel distributions) have logarithmic circuit depth as a function of file size?
@robryk I know that, but your statement should be amended to "whenever it is required". It would already be quite useful even for normal O(MB) file sizes to be able to use parallelism, which is why I am asking about the default hash.