Re: Noise generator in VHDL
in addition to the stuff from bartart:
xapp052 (especially this one), xapp217, xapp220 from Xilinx documentation.
These are LFSR (Linear Feedback Shift Register) designs that produce pseudo (careful!)-random sequences. This means that eventually they get repeated. But if you have e.g. a 32-bit LFSR producing a maximal length sequence, they won't repeat for 4 billion or so cycles.
I have tried them and they synthesize to smallish hardware. And they are fast too.
For better quality generators, you could sample the thermal noise of a resistor (or a pn-junction maybe) and use an ADC to get completely uncorrelated random values.
I'm not sure that the pn-junction will give uniform distribution though (i doubt that, but too lazy to get the maths done
LFSRs have zillionths of uses (CDMA, PRNGs for roll-dice problems, in game engines etc).
*** FUN ***
I love playing that old soccer game "KickOff3". It is far better than its famous predecessors. I have noticed that the goalkeeper's behavior is rather intelligent for such an old dog!!!
Let me be more specific. The goalkeeper doesn't always make the same attempts for blocking an opponent shot! In all the other oldies, a simple algorithm is used so that the goalkeeper is in a certain relative position/angle to the ball. Which eventually means, that he ALWAYS saves specific shots, and ALWAYS gets beat by specific shots depending on the position/angle! In KickOff3 THIS IS NOT THE CASE!!! I believe they use some cheap PRNG in hardware, so that he seems to have "randomish" behavior!!!
Anybody knows about this?
cheers
the_penetrator_in_extremis©