I wanted to make sure it captured some non-max gaps since that seemed to be what people were interested in.

[QUOTE=Antonio;459743]Also note the low gap (=1100) that was used, that would slow the process quite a bit, if he had used 1248 it would have been faster, and a fairer comparison with other peoples speeds.[/QUOTE]

Also see the low sb value, for that xeon the optimal could be sb=27 or sb=28.
But even with that it is possible that you can reach the memory bandwidth using only t~10 threads.

4211-4218e15 complete
[CODE]4210-4211e15 93 kgaps, largest 1176 @ 4210279718438929981
4211-4212e15 96 kgaps, largest 1212 @ 4211308991831021999
4212-4213e15 106 kgaps, largest 1224 @ 4212891798760396469
4213-4214e15 92 kgaps, largest 1224 @ 4213835242980370009
4214-4215e15 146 kgaps, largest 1212 @ 4214607784623404249
4215-4216e15 112 kgaps, largest 1184 @ 4215348541272003419
4216-4217e15 123 kgaps, largest 1176 @ 4216415379215364287
4217-4218e15 95 kgaps, largest 1152 @ 4217090183575843837[/CODE]
Reserving 4220-4230e15

And so we have now done 100 ranges (10%) with a further 60 ranges reserved. It took 52 days from the start of the project, some of which was spent writing and experimenting with code, and recruiting searchers. However the next 100 ranges should be completed much quicker. Some statistics:

[CODE]

searcher compl. res'd

Thomas11 32 8
antonio 31 10
robert44444uk 29 16
pinhodecarlos 6 4
danaj 1 20
ET_ 1 2

Total 100 60

[/CODE]

This morning I got my first merit 30 gap from the search - a gap of 1306

4132e15 - 4140e15 complete:

[CODE]
A B C D E F G

4132 4133 Thomas11 completed no 1252 4132321296601727707
4133 4134 Thomas11 completed no 1268 4133768580472197581
4134 4135 Thomas11 completed no 1180 4134342756975622069
4135 4136 Thomas11 completed no 1194 4135948685233934773
4136 4137 Thomas11 completed no 1322 4136179975357722581
4137 4138 Thomas11 completed no 1214 4137136145310653777
4138 4139 Thomas11 completed no 1152 4138265290485105667
4139 4140 Thomas11 completed no 1198 4139428867196431453
[/CODE]

Reserving 4140e15 - 4160e15.

4220e15 - 4225e15 complete, no new records.[CODE]
4220-4221e15 99 kgaps, largest 1184 @ 4220838055436693423
4221-4222e15 69 kgaps, largest 1210 @ 4221700962773095519
4222-4223e15 98 kgaps, largest 1208 @ 4222726409250843239
4223-4224e15 103 kgaps, largest 1362 @ 4223823832453807411
4224-4225e15 93 kgaps, largest 1164 @ 4224425843133096203
[/CODE]

[QUOTE]4223-4224e15 103 kgaps, largest 1362 @ 4223823832453807411[/QUOTE]This gap of 1362 after 4223823832453807411 is the first one ≥ 1346 that has not been previously discovered by LLh.

A 1362 gap has a First Occurrence at 3937457950646269397

[QUOTE=rudy235;459967]This gap of 1362 after 4223823832453807411 is the first one ≥ 1346 that has not been previously discovered by LLh.

A 1362 gap has a First Occurrence at 3937457950646269397[/QUOTE]

Not quite right. LLh may have discovered this 1362 at 42238... There are no published records for second instance prime gaps so unless LLh lets us know he did not find that gap we will never know if your statement is true.

It is really difficult to reword your remark in a way which is true. Maybe:
[I]
This gap of 1362 after 4223823832453807411 is the smallest prime gap ≥ 1346 discovered to date by the Mesennefourm 4e18-5e18 group that is (a) greater than 4e18, (b) less than 4.225e18 and (c) that is not a first instance prime gap.[/I]

Which is not particularly exciting.

In ref. to the previous post.
Well that was what I, perhaps ineptly, I tried to convey. However I did not want to make it a complicated statement.

Here we go.

In 2011 Tomas Oliveira e Silva finds the gap of 1362 at 3.937 E18

At that time comprehensive analysis of gaps had been done up to 2.6 E 18

Because of that, the gap was only confirmed to be a "first occurrence" when Tomas Oliveira e Silva (TOeS) finally completed an exhaustive check up to 4.0 E18 on April 4 2012.

In 2014 when ( again we presume ) LLh was checking the range 4.2 -- 4.3 E18 he would have found the gap and would have been aware of the 1362 gap stating at 3.937 E 18 which had already been discovered 3 years prior and was since April 4th 2012 a confirmed "First Occurrence".
This would be consistent with LLh finding the gap and NOT reporting it.

However --I agree-- this is all speculation or at the best approximation to the truth. The only thing we can say with certainty is that up to today no gap >= 1346 that has been found in this 2017 search is a "first know occurrence".

[QUOTE=rudy235;459991] The only thing we can say with certainty is that up to today no gap >= 1346 that has been found in this 2017 search is a "first know occurrence".[/QUOTE]

That's how I would read it :cool:

It will be nice to get into the 4.6e18 range where I reckon LLh possibly stopped. I think maybe he took it up to 2^62 = 4.611e18 as we have a big gap between there and the next smallest early occurrence gap.

I don't think it will take us too long to get there either - possibly 2 months if we can process 1e11 per second. That's assuming we have five contributors capable of 2e10 per sec. We have 3 maybe 4 contributors at that rate right now.

LLh went to at least 4.6460 e18 in 2015. I.e. Gap of 1346
I have the feeling he did not go past 4.7 e18, but he is not sharing this info.:nocomment: