Cheating with low voltages on a DK board

I never said any such thing.
Please cite where I said that. I'll wait.

That said, it's technically possible (and it's been done) where random number generation can be done using analog elements, which could be voltage-sensitive. But that isn't typically what you see in digital systems.
You could have just scrolled back.

You're really just demonstrating a lack of understanding of the difference between RNGs and PRNGs. Having done work in both communications and encryption -- it matters.
 
8bd81aaf-d811-4e43-8614-52cd02febf40_text.gif
 
You could have just scrolled back.

You're really just demonstrating a lack of understanding of the difference between RNGs and PRNGs. Having done work in both communications and encryption -- it matters.


And if you actually read what I said, I sad 'not typically', meaning they aren't as frequent. Not that they don't exist at all. Do I need to explain how to understand English again to you?

In any given computing device these days, how many analog-based RNG's vs digital PRNG's are there? I'm also including hardware AND software, which is part of 'digital systems', not that it changes the answer.

And if you read my other posts, you'd see that I'm fully aware of the difference, and when one is needed over another. And for a vast majority of cases today, PRNG is sufficient. That doesn't mean I said entropy wasn't used anywhere, or that it wasn't important, or that there aren't applications where it's critical (like cryptography). Just that the NUMBER of places it's needed is tiny compared to PRNG. Stop with the strawmanning.

Can you stop with the dick flexing already? I worked in communications too. (And I do again now.) Either you're unintentionally misinterpreting what I said, or you're intentionally doing it. But either way you have no case.
 
And if you actually read what I said, I sad 'not typically', meaning they aren't as frequent. Not that they don't exist at all. Do I need to explain how to understand English again to you?

...and for RNGs, digital almost never used since they're not actually random... so yes... analog is "typically" (ie. almost always) used.

In any given computing device these days, how many analog-based RNG's vs digital PRNG's are there? I'm also including hardware AND software, which is part of 'digital systems', not that it changes the answer.

Your statement was regarding RNGs, not PRNGs, but sure -- keep moving the goalposts.
 
...and for RNGs, digital almost never used since they're not actually random... so yes... analog is "typically" (ie. almost always) used.

Your statement was regarding RNGs, not PRNGs, but sure -- keep moving the goalposts.

I'm not moving any goalposts. (And you're projecting, BTW.) It isn't my fault you're trying to twist what I'm saying, but nice try. Again, another example of you being overly literal (and pedantic) as a way of arguing. Let's review:

When I said RNG (especially in the context of THIS DISCUSSION, which has to do with 40-year-old boards) I obviously NOT referring to entropy, and was using the term in the general sense to refer to whatever system (or method) the board uses to generate the randomness needed. (Which in this case will be digital, and hence technically pseudorandom).

I was not saying it had to be a discrete circuit. I did not say if it had to be hardware or software. If you read post #9, I explicitly say 'random number generation' and then use the acronym RNG, *without* specifying an implementation.

So are you trying to argue now that 'RNG' specifically has to mean entropy? Horseshit. You yourself used the term in this thread when referring to Pokeys (which are a 17-bit LFSR, and thus technically a PRNG, kiddo):

Bad random number generator on the Pokey...

It's socketed, so you can just pull it and drop a new one in.

I keep a centipede board with a bad RNG but good sound around to swap in to make it centipede for dummies for when incompetent players visit....

Are you done now?
 
There is no "random number generator" on the PCB. Donkey Kong doesn't even use the DRAM refresh register for pseudo-random numbers. It's just counters being updated during the ISR.

Wouldn't slowing the counters also slow everything else going on in the game, such as mario's (jumpman) movement through the girders, ladders around the screen? The original claim was it would only affect the fireballs, which seems pretty selective.
Any visually perseptual change would have to be large in digital terms (voltage sag, duty cycle time) to even be noticed. I haven't seen the DK code to know how many lines the loop is for the "fireball movement", but I imagine the time it would take to loop through it would be in microseconds? You would have to pulse the low voltage at just the right time and for the correct duration to only affect the that section of code.
I'm basing this on Mark pointing out there are no random number generators built in TTL circutry.

Mark, do you think there would be any hysteresis effect in the flops going into the NMI by holding the rail voltage at a low power threshold? I'm betting it would be neglegible but just tossing it out there.
 
Wouldn't slowing the counters also slow everything else going on in the game, such as mario's (jumpman) movement through the girders, ladders around the screen? The original claim was it would only affect the fireballs, which seems pretty selective.
Any visually perseptual change would have to be large in digital terms (voltage sag, duty cycle time) to even be noticed. I haven't seen the DK code to know how many lines the loop is for the "fireball movement", but I imagine the time it would take to loop through it would be in microseconds? You would have to pulse the low voltage at just the right time and for the correct duration to only affect the that section of code.
I'm basing this on Mark pointing out there are no random number generators built in TTL circutry.

Mark, do you think there would be any hysteresis effect in the flops going into the NMI by holding the rail voltage at a low power threshold? I'm betting it would be neglegible but just tossing it out there.
unfortunately there is no possible way to slow down the "fireball" part of the code without slowing down everything. The CPU is a sequential machine so if the clock were to slow during the "fireball" portion of the code ALL of the code for the rest of the game would then be delayed by that amount. The clock is generated with a crystal so large voltage variation will only change it very small amounts ~100ppm or less (.01%) that's not going enough to perceive a difference. The clock also generates the video and the sync pulse for the monitor, if it changes more that a few percent the monitor would have difficulty locking to the sync. If by hysteresis you are talking about propagation delay change due to voltage variation that does exist BUT it's a few nanoseconds at the most and it doesn't accumulate. It will be seen as jitter on the clock but won't change the frequency of the clock.
 
unfortunately there is no possible way to slow down the "fireball" part of the code without slowing down everything.
Well, pretty straightforward to do by hacking the ROMs.... but not by playing with voltages.

Pretty stupid that all TG cares about is ROM labels, especially on Nintendo where the labels don't even cover the window, so it'd be trivial to erase and reprogram and still have them "look" original.
 
Well, pretty straightforward to do by hacking the ROMs.... but not by playing with voltages.

Pretty stupid that all TG cares about is ROM labels, especially on Nintendo where the labels don't even cover the window, so it'd be trivial to erase and reprogram and still have them "look" original.
Yes, my comment was relating to the "voltage" variation discussed in the thread. Totally agree with you on how TG "regulates" or in some cases fails to "regulate" key parameters like validating roms/hardware. I don't think they have the technical resources available to them to really do a good job in that regard.
 
This thread should have been closed a long time ago.

Now more than ever i believe this thread needs to be closed, sterilized and nuked from orbit.


I really hope it does not get deleted, like other threads have.

Aside from the furry shit, there's valid information here that people who aren't familiar with can learn about. Folks can google random number generation, and learn about the different types that exist, the strengths, weaknesses, and costs of each, and when some are needed over others. You can also google statistical analysis of random numbers, measures of randomness, and other related topics. Then everyone might have a better ability to understand the issue in the OP, and how it could be investigated.

Or people can shitpost.
 
Back
Top Bottom