Efficient Random Number Generation

[shawn]

You could use if to seed a boat load of other sic's.

[oar_some]

use the provided functions?!

[jonathan]

use the provided functions?!

What provided functions, where?

Not helpful.

[that_awesome_guy]

um...

Code: Select all

#include <xs1.h>
#include <math.h>
#include <print.h>
#include <stdlib.h>

int main()
{
	printuintln(rand());
	return 0;
}

:?: :?: :?: :?: :?: :?: :?: :?:

[jonathan]

Thanks, but the sarcasm wasn't needed. Where are these documented?

Don't you work for XMOS? I thought XMOS staff were supposed to be helpful! It feels like you are trying to make me out to be an idiot. That isn't what this site is about. Or so I thought?

[that_awesome_guy]

I thought it was a helpful pointer, sorry you didn't feel that way.

These are standard C functions, they are not xmos specific. If XMOS supports standard C then I would personally expect it to support standard library functions such as these. Standard library documentation can be found here - http://en.wikipedia.org/wiki/Stdlib.h

[jonathan]

Thank you. Much more helpful and not in the least bit patronizing. :-)

[dan]

The L1 series has some ring oscillators which were put in for our characterisation purposes. They are free running counters that have no relationship with any other clocks and oscillate with a PVT dependant frequency between 300 and 700 mhz, if you enable them.

They can be read in software with a GETPS, but are not documented AFAIK.

Would these be of any use?

[otitov]

anyway it would be very interesting to try them!

[shawn]

The L1 series has some ring oscillators which were put in for our characterisation purposes. They are free running counters that have no relationship with any other clocks and oscillate with a PVT dependant frequency between 300 and 700 mhz, if you enable them.

They can be read in software with a GETPS, but are not documented AFAI?

This is interesting, I wonder what other eggs we may find. perhaps useful for simulation.