[DannyBee]

That's the problem though. Every report of speed increase is unscientific, and every scientific mechanism shows it can't possibly speed anything up.

nothing is opening and using /dev/random continuously, as shown by using inotify to watch it.

Things are using /dev/urandom, but /dev/urandom doesn't block anyway.

I see a bunch of people showing that this patch makes more entropy available (duh), but without anything using /dev/random, that should not matter.

The only thing i can think of is that something in the kernel is asking for random numbers from the blocking interface. Otherwise, it should show up in inotify.

[vidarh]

This assertion:

> and every scientific mechanism shows it can't possibly speed anything up.

Does not follow from this:

> nothing is opening and using /dev/random continuously, as shown by using inotify to watch it.

The finding that nothing is opening and using /dev/random rules out the explanation the author of the app gave. It does not in any way prove that the app has no effect, even if it might be for totally different reasons.

Several people appear to be looking at ways of doing some proper tests to determine if there is a real measurable effect, so we'll presumably find out soon enough.

[BenoitEssiambre]

It is also possible that this has nothing to do with blocking but rather with processor resources depletion. The fact that the entropy generating processes are being triggered to run at just the wrong time when the device is most likely busy instead of being triggered when it is idle might cause slowdowns.