GAPS BETWEEN PRIME PAIRS (Twin Primes)

[firejuggler]

So, we are looking at gap >3103?

[robert44444uk]

Originally Posted by firejuggler

So, we are looking at gap >3103?

There are numerous gaps above 3103 that exist in the space up to 1.2e15, but 3103 is the smallest gap not in that space.

I'm exploring 1.2e15 to 9e15 using Thomas11's suite of programmes.

[firejuggler]

So, how can I help? currently looking at above 1.5e16,

Since we have found everything upto 3103, the only interesting things are above, right?


like

3169 15000007142666023

[robert44444uk]

Originally Posted by firejuggler

So, how can I help? currently looking at above 1.5e16,

Since we have found everything upto 3103, the only interesting things are above, right?


like

3169 15000007142666023

Correct...but

Fischer already found 3169 148261013820413

The best way to approach this is to just run with Thomas11's suite and start at 9e15 - your results over a large range will then feed into the broader picture when I eventually catch up. If you want to start at 1.5e16 that's fine, but just post everything > 3103 once you get to, say 1.55e16 or 1.6e16.

[firejuggler]

I just wanted to be sure I understood what I should do. I will start at 9e15 and go to 9.05e15 to see the scale/ lenght of the project


edit: are we missing 3016?

[robert44444uk]

Originally Posted by firejuggler

I just wanted to be sure I understood what I should do. I will start at 9e15 and go to 9.05e15 to see the scale/ lenght of the project


edit: are we missing 3016?

You can assume that. I'm not at home to check. Welcome to the search!

[Thomas11]

Originally Posted by firejuggler

So, how can I help? currently looking at above 1.5e16,

Since we have found everything upto 3103, the only interesting things are above, right?


like

3169 15000007142666023

Originally Posted by robert44444uk

Correct...but

Fischer already found 3169 148261013820413

The best way to approach this is to just run with Thomas11's suite and start at 9e15 - your results over a large range will then feed into the broader picture when I eventually catch up. If you want to start at 1.5e16 that's fine, but just post everything > 3103 once you get to, say 1.55e16 or 1.6e16.

Just to make sure that we all use the same "language" when refering to gaps and ranges:

Fischer ran his search up to "1.28e15" in "real" numbers.
When Rudy235 started this thread, he was looking at numbers of the form 6*k+-1. The script/NewPGen combo also uses this form. Thus, Fischer's limit of 1.28e15 translates to 2.1333e14 in terms of k.

The same holds for the gap lengths: Our gaps are 1/6th of Fischer's gaps.

So, in order to continue from Fischer's upper bound we need to start from k=2.1333e14.

[Thomas11]

To give you some idea for planning your work (ranges):

The Perl/NewPGen combo is able to crunch a range of about 1e11 (in terms of k) per hour on a single cpu core.

A range like Fischer's 1.28e15 (=6*2.1333e14) would take about 90 (single core) cpu days. Extending Fischer's range up to 2.0e15 (=6*3.3333e14) would take 50 cpu days.

[firejuggler]

Then I searched 6 time higher than I intended.if I should search at k= 15e6 for real 9e15 and beyond, right?

[robert44444uk]

Sorry - my bad. Thomas11 is right about the situation. In his run_new.bat file, you need to set kmin at 9000000 and kmax at say 9500000

I have 14 cores working on this starting at kmin 2333333 up to 8000000

[robert44444uk]

The best merits to date are all Fischer's

4205	191569830235030	Fischer	27.70057197
3893	123819102042695	Fischer	26.30315806
3758	75135208923870	Fischer	26.14839171
3882	132146379263218	Fischer	26.12940501
3931	192276915343362	Fischer	25.89008676
3655	58995174450558	Fischer	25.80021286
3897	186964141272992	Fischer	25.70768054
3703	90551918956495	Fischer	25.48301549
3713	101014339594590	Fischer	25.38793801
3733	124779667215570	Fischer	25.21073307