Closed digitalsignalperson closed 1 day ago
Hi!
It's not particularly meant to be fast. I really just wanted that tool to be compact, simple and reliable.
Maybe I'll do a super fast variant using libaegis, but in the meantime, encpipe has been protecting my backups for over 7 years, and for that kind of use case, performance was not the priority.
Yeah totally appreciate the compact/simple/reliable. And cool to see libaegis as an option.
I'm also curious about the gimili permutation as something being simple but also potentially fast... Gimli Encryption in 715.9 psec sounds like it has potential for competitive performance (though they mainly talk about latency not throughput, not sure how to interpret). Any thoughts on acceleration potential?
That paper is about hardware implementations; i.e. putting Gimli directly in a chip.
ah oops, i feel dumb for scanning it and missing the point. thanks for the replies, I like the tool
Hey I was curious what performance to expect with this tool.
On a beefy PC I'm getting ~37MiB/sec with a
dd if=/dev/zero bs=1M count=1000 oflag=direct | encpipe -e -p testpass > /dev/null
Is it a limitation of libhydrogen or the gimli implementation?