calcyman wrote: ↑August 1st, 2010, 1:22 pm
Nope, it's Moore. The diagonals are necessary for pretty much everything.
Well, that changes things completely.
Complexity of Wave CA = (ln(12)/ln(2))*(12^9) ~ 1.85 * 10^10 bits
Complexity of vNCA = (ln(29)/ln(2))*(29^5) ~ 9.96 * 10^7 bits
So, Wave CA is more complex than von Neumann's CA.
I /think/ Hutton32 isn't able to do a logical NOT at all, since it can't universally destruct unlike the original JvN29 or Nobili32.
It can still destruct OTS cells, which is sufficient for logical negation.
That's quite ridiculous looking, I must say.
Yes, it's an AND-NOT coupled with a XOR gate. And the AND-NOT gate involves a crossover, which is another three XOR gates.
What exactly do you mean by a latch? If you want an R/S flipflop
Similar to an R/S flipflop, but more passive. It has two inputs, FLIP and TEST, and two outputs (ON and OFF), like so:
Code: Select all
x = 50, y = 21, rule = Nobili32
29.L$29.J$29.J$29.J$29.J$29.J$29.J$29.J$29.J$27.Q.J$IM15IM9I2pA4K2O
13KOK$27.QIL$29.L$29.L$29.L$29.L$29.L$29.L$29.L$29.L$29.J!
The set {fanout, latch, merge} is universal, as well as the more conventional set {fanout, NAND}. The advantage of the first one is that there is no synchronisation involved.
I'm sure you could do it in 128x128 with Nobili32.
I think I could manage a 32*32 Unit Life Cell in Nobili32. Here is the logic circuit:
Code: Select all
x = 28, y = 20, rule = Nobili32
5.5IL.IL$5.JL3KLI2pA2IpA2IpA2IL$5.J2ILpAKJ2pA2KpA2KpA2KpA$5.J2.4IJL2.
L2.L2.L$5.J2IL3.TpA2TpA2TpA2TpA$5.2JIpAIL.10TL$5.3JIpA2I10.L$2.3I3J2K
J.2J2L5J.L$2IpA3IpA3IpA.10IpAL$2.J3.J3IL.3LJ5L.2L$2.J3.2JpAK2I10.2L$
2.JL2KpA.JpAJ.10R2L$2.JLQ.J2.J2.RpA2RpA2RpA2R2pA$2.JI2pA2K.J3.J2.J2.J
2.2L$2.2JQ2IpAIJ3.pA2KpA2KpA3KL$2.2J3.J15.L$2.2J3KpA15KpA4I$2.J4.J$2.
J4KpA$7.J!
The pulses from the 8 neighbours enter from the left. The signal from the right should be directed to all neighbours. Periodically, a signal from the bottom updates the cell.