True Random Number Gen in Virtex 7

[Kevin Neilson]

Has anybody tried to build a true random number generator from Catalin Baetoniu's paper?

http://forums.xilinx.com/xlnx/attachments/xlnx/EDK/27322/1/HighSpeedTrueRandomNumberGeneratorsinXilinxFPGAs.pdf

He uses interconnected ring oscillators in conjunction with a "Linear Hybrid Cellular Automata" for making the distribution more uniform.

I built this in a Virtex 7 to use in the lab for hardware testing. I had to instantiate the LUTs for the ring oscillator and put DONT_TOUCHes on them.

I just wondered if anyone else had tried this. It's difficult to tell "how random" the numbers really are without extensive analysis. I noticed Xilinx has a patent based on this paper, but I didn't know you could patent ring oscillators.

 

[Tim Wescott]

On Tue, 20 Dec 2016 18:44:08 -0800, Kevin Neilson wrote:

Has anybody tried to build a true random number generator from Catalin
Baetoniu's paper?

http://forums.xilinx.com/xlnx/attachments/xlnx/EDK/27322/1/
HighSpeedTrueRandomNumberGeneratorsinXilinxFPGAs.pdf

He uses interconnected ring oscillators in conjunction with a "Linear
Hybrid Cellular Automata" for making the distribution more uniform.

I built this in a Virtex 7 to use in the lab for hardware testing. I
had to instantiate the LUTs for the ring oscillator and put DONT_TOUCHes
on them.

I just wondered if anyone else had tried this. It's difficult to tell
"how random" the numbers really are without extensive analysis. I
noticed Xilinx has a patent based on this paper, but I didn't know you
could patent ring oscillators.

Ring oscillators to generate -- or at least seed -- random numbers isn't
new. If there's a patent there, they're probably trying to make a real
or perceived barrier to people doing it with an Altera part.

--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com

 

Den lørdag den 24. december 2016 kl. 21.48.55 UTC+1 skrev Tim Wescott:

On Tue, 20 Dec 2016 18:44:08 -0800, Kevin Neilson wrote:

Has anybody tried to build a true random number generator from Catalin
Baetoniu's paper?

http://forums.xilinx.com/xlnx/attachments/xlnx/EDK/27322/1/
HighSpeedTrueRandomNumberGeneratorsinXilinxFPGAs.pdf

He uses interconnected ring oscillators in conjunction with a "Linear
Hybrid Cellular Automata" for making the distribution more uniform.

I built this in a Virtex 7 to use in the lab for hardware testing. I
had to instantiate the LUTs for the ring oscillator and put DONT_TOUCHes
on them.

I just wondered if anyone else had tried this. It's difficult to tell
"how random" the numbers really are without extensive analysis. I
noticed Xilinx has a patent based on this paper, but I didn't know you
could patent ring oscillators.

Ring oscillators to generate -- or at least seed -- random numbers isn't
new. If there's a patent there, they're probably trying to make a real
or perceived barrier to people doing it with an Altera part.

afaict patents have become something big companies have as ammunition against small guys: try to enter the market and will bury you in lawsuits
and protection against other big companies: don't sue us will sue you right back and we'll both waste money

 

[Tim Wescott]

On Mon, 26 Dec 2016 15:44:02 -0800, lasselangwadtchristensen wrote:

Den lørdag den 24. december 2016 kl. 21.48.55 UTC+1 skrev Tim Wescott:
On Tue, 20 Dec 2016 18:44:08 -0800, Kevin Neilson wrote:

Has anybody tried to build a true random number generator from
Catalin Baetoniu's paper?

http://forums.xilinx.com/xlnx/attachments/xlnx/EDK/27322/1/
HighSpeedTrueRandomNumberGeneratorsinXilinxFPGAs.pdf

He uses interconnected ring oscillators in conjunction with a "Linear
Hybrid Cellular Automata" for making the distribution more uniform.

I built this in a Virtex 7 to use in the lab for hardware testing. I
had to instantiate the LUTs for the ring oscillator and put
DONT_TOUCHes on them.

I just wondered if anyone else had tried this. It's difficult to
tell "how random" the numbers really are without extensive analysis.
I noticed Xilinx has a patent based on this paper, but I didn't know
you could patent ring oscillators.

Ring oscillators to generate -- or at least seed -- random numbers
isn't new. If there's a patent there, they're probably trying to make
a real or perceived barrier to people doing it with an Altera part.

afaict patents have become something big companies have as ammunition
against small guys: try to enter the market and will bury you in
lawsuits and protection against other big companies: don't sue us will
sue you right back and we'll both waste money

Yes, the USPTO has become something of a captive regulator to the big-
money interests.

--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com