Author Topic: Cryptography Implementation Discussion  (Read 18187 times)

Making it asynchronous via schedules is essentially bullstuffting the "doesn't lag BL" factor.

If you divvy the work up small enough, no, it isn't.

I didn't say that it would still make BL lag, it wouldn't. But you're basically lying by saying that it doesn't lag BL, instead of making it lag BL you're making it lag the calculations. The calculations in series would still lag Blockland, which is why I'm saying any solid cryptography implementation would cause enough stress to lag Blockland. It's obvious that if you stagger it among hundreds of ticks it won't lag, but that's more like a cheap hack to make me wrong by technicality. You could run 8192 bit RSA through Blockland if you really forgetin' wanted to without any notable lag as long as you make it asynchronous. It'll just take an eternity.

Yeah, a cheap hack if you must, but it's still going to work. No one has ever gotten it working before. It doesn't have to be super fast, slow is fine as long as it isn't taking unreasonably long, like ten to twenty minutes for anything other than, say, key generation, and that's still pretty slow for that.

And I'm not trying to make you wrong, but you can't casually brute force this stuff that easy. I simply don't believe it.

We don't have unreal expectations about speed or anything. I have personal hopes about the speed, but we as a group are simply trying to get it working as well as we can. The general consensus is, or at least was, that you can't get high level crypto working in torque. We aim to prove that wrong.

Trinick, just drop it. If everyone assumed that something can't be done, then it never will be. You can believe it's impossible if you want, just don't waste the time of the people who do. If you're right, big deal. It's still their choice to try, not yours.

I just implemented that exp function on my calculator

It's super loving fast. Like, less than a quarter of a second for numbers like 529474^399271 % 247
I think the calculator overflows after that

I can't casually brute force actual high grade encryption, you're right. I was under the assumption you're simplifying cryptographic algorithms so that they could be run laglessly via Blockland, not just running existing ones asynchronously.

Goodness no, I'm not dumb enough to spin my own crypto.

Haha no. At the moment I'm looking at other ways to calculate modulus, becuase that's incredibly slow at the moment.

Whenever you get that working, implement the exponential modulus thingy. Here's some pseudo code typed on my phone

Exp (x, e, m)
Y = 1
Z = X
While e greater than zero
If e is odd, then
Y = (y times z) mod m
(End if)
Z = (z times z) mod m
E = Divide and round down e by 2
End while loop
Return y

I dunno why it works, but it does.

Yeah I already got the pseudocode from the python code for it that you posted earlier. I'll implement that as well.

Eh, didn't know if you were successful or not.

Haha no. At the moment I'm looking at other ways to calculate modulus, becuase that's incredibly slow at the moment.

Yeah sorry I didn't get to the modulus earlier. I haven't had an ounce of free time all week really.

Goodness no, I'm not dumb enough to spin my own crypto.
Dumb enough?

Pretty sure that people who write cryptographical algorithms are highly intelligent.