Limiting the consecutive occurrence

[M@LIK]

Edit: I got it working but can't be sure as I don't understand the command at all but notice you added

Code:

/\(.\).*\1.*\(.\).*\2/d

and

Code:

/\(.\).*\(.\).*\1.*\2/d

so I remove one of them. The command I using that seems to work is

Code:

sed "/\(.\)\1\1/d;/\(.\).*\1.*\1/d;/\(.\).*\1.*\(.\).*\2/d"

and output is

Code:

AABCDEFG
AABCDEFH
AABCDEFI
AABCDEFJ
AABCDEFK
AABCDEFL

these extra commands have slowed it down even more lol, shame they is no GPU version of sed or even a mulit-core version

I added those for the rule you wanted, I don't know how you got it working without them.
And yes, the more commands, the slower.

Just had another idea for sed, can it also delete lines that have any 6 character alphabet sequence within them?

Yes, of course!
But, there's no need to do that, it will be slower.
Just use "-s AABCDEGA" to skip all those candidates as they won't appear again.

how large is one file generated by such a command?

Huge! Hash-IT generated all the possibilities beginning with "A" without any filtration, it was 75GB, so 75x26=1950GB, let's say we can filter 25% = 487.5GB
I barely have 10gb free on my 2tb xD
Anyways, we're still working on the filter, so nothing to start with.