[sdiy] Non maximal-length LFSR

[Tom Wiltshire]

This sounds like what I tried:

http://electricdruid.net/practical-lfsr-random-number-generators/

The gist of it is that you can (for certain tap sequences) process eight bit shifts at a time and thereby do the XORs in full bytes and the bit shifts as a (much more efficient) byte shift instead.

Tom

On 4 Mar 2016, at 19:45, Eric Brombaugh <ebrombaugh1 at cox.net> wrote:

> I've done this on DVB modems that require byte-wide scrambling. I started with the raw single-bit LFSR specification and then just did the boolean logic required to compute 8 bits per cycle. Instead of shifting one bit in the register you shift 8 which means you need to compute 8 unique inputs at a time. More XOR operations at once.
> 
> I suspect that if your dsPIC implementation is having trouble computing a 2-tap LFSR then this might require more logic than you care to do.
> 
> Eric
> 
> 
> On 03/04/2016 12:26 PM, Richie Burnett wrote:
>> How is this done Dave?  Do you have a reference?
>> 
>> -Richie,
>> 
>> -----Original Message----- From: Dave Manley
>> Sent: Friday, March 04, 2016 7:07 PM
>> To: Richie Burnett ; Tom Wiltshire
>> Cc: synth-diy DIY
>> Subject: Re: [sdiy] Non maximal-length LFSR
>> 
>> It hasn't been mentioned, but the LFSR implementation can be done so
>> that it generates multiple bits per update, not just a single bit.  This
>> is commonly done in telecom/ networking descramblers which operate on a
>> multi-bit serdes output.
>> 
>> -Dave
>> 
>> On March 4, 2016 5:24:02 AM PST, Richie Burnett
>> <rburnett at richieburnett.co.uk> wrote:
>>> Yeah, but throwing away every 49th result would be throwing away 7
>>> years
>>> worth! ;-))  Could I live with that!?!?
>>> 
>>> Thanks for everyone's help and suggestions about this.  I'm confident I
>>> know
>>> the way forward now, and also have some good pointers towards working
>>> through the maths to back up my decisions, which always makes me feel
>>> better.
>>> 
>>> -Richie,
>>> 
>>> 
>>> 
>>> -----Original Message----- From: Tom Wiltshire
>>> Sent: Friday, March 04, 2016 12:17 PM
>>> To: Richie Burnett
>>> Cc: mskala at ansuz.sooke.bc.ca ; synth-diy DIY
>>> Subject: Re: [sdiy] Non maximal-length LFSR
>>> 
>>> 371 years!!
>>> 
>>> On 4 Mar 2016, at 11:34, Richie Burnett <rburnett at richieburnett.co.uk>
>>> wrote:
>>> 
>>>> Sure the resulting sequence will be a smidgen shorter, but 2^49-1 is
>>> quite
>>>> long anyway.
>>> 
>>> 
>>> 
>>> -----
>>> No virus found in this message.
>>> Checked by AVG - www.avg.com
>>> Version: 2016.0.7442 / Virus Database: 4537/11745 - Release Date:
>>> 03/04/16
>>> 
>>> _______________________________________________
>>> Synth-diy mailing list
>>> Synth-diy at dropmix.xs4all.nl
>>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>> 
>> 
>> 
>> -----
>> No virus found in this message.
>> Checked by AVG - www.avg.com
>> Version: 2016.0.7442 / Virus Database: 4537/11746 - Release Date: 03/04/16
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>> 
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy