Very Prime Riesel and Sierpinski k

[pinhodecarlos]

Thomas, I won't do nothing. Files will stay as they are. Thank you.

[Thomas11]

Based on Carlos' question I did a little statistics on all the Sierpinski sequences I processed so far. Please find the data attached.

The files will look like the following example (for E=66):

level min max average
   50   5  18  10.10
  100  12  27  17.44
  200  20  39  27.52
  300  28  45  34.57
  500  36  54  44.47
  750  45  65  52.80
 1000  53  69  58.98
 1500  59  79  67.10
 2000  65  83  73.01
 2500  69  90  77.73
 3000  75  96  81.59
 4000  79 104  87.64
 5000  85 105  92.43
 6000  91 106  96.26
 7000  92 109  99.24
 8000  94 110 102.00
 9000  99 113 104.36
10000 105 115 106.60
data from 429 files

Given are the minimum and maximum number of primes at the given level, as well as the average over all sequences for the specific E.

Note that the found "min" values more of less replicate the actual Smith check levels used in my in.txt files.

Based on that data one may optimize the levels by setting them to something like:
count(level) = average(level) - 6

But, perhaps, Robert should decide on this...

Perhaps we could find some (simple) functional dependency (also depending on the E level), which could replace the hard set Smith check levels in a future version of the software. Maybe with an adjustment parameter for the "stiffness"...

[robert44444uk]

Really busy today, but my computer goes on regardless :)

Nice result for Riesel, into virgin territory, breaking a record that has stood for many years:

160 143571 440310850049907 R 82
161 158755 440310850049907 R 82
162 159331 440310850049907 R 82
163 160412 440310850049907 R 82

Concentrating now on the upper Riesel ranges: 180, 196, 210, 226

[robert44444uk]

Originally Posted by Thomas11

But, perhaps, Robert should decide on this...

Perhaps we could find some (simple) functional dependency (also depending on the E level), which could replace the hard set Smith check levels in a future version of the software. Maybe with an adjustment parameter for the "stiffness"...

Every man can decide his own targets, getting 100/10000 is not noticed really unless your target is most VPS.

The system has "stiffness" already built in

c0 1.5
c1 5.0

These define the stiffness. Yon should trawl back to Robert G's original postings to learn about this.

[pinhodecarlos]

My weeks results. Pay attention for duplicates because I set I=0 when I updated the client to the new one.

Carlos

[pinhodecarlos]

Thomas,

About the table you posted here, did you do it manually or do you have some kind of SQL database? How hard is to make one table online where you could upload your results.txt files and automatically get that stats table?

Carlos

[robert44444uk]

Working in the upper echelons is still not easy work, the iterations fly by with little evidence of any very prime candidates.

But I have one Riesel E180 VPS

R 3520085628118353 180 102/10000

And on near miss at E196

R 1624082107451859 196 99/10000

[Thomas11]

Originally Posted by pinhodecarlos

Thomas,

About the table you posted here, did you do it manually or do you have some kind of SQL database? How hard is to make one table online where you could upload your results.txt files and automatically get that stats table?

Carlos

Carlos,

I wrote some Perl scripts for generating this kind of tables (see the attachment).
But I have no experience in how to use them in a Windows environment...

Thomas

[Thomas11]

During the weekend I found the first VPS for E=196:

S 12371760550032321 196 100/8794 102/10000

And, even more exciting, the first E=292 Payam sequence:

S 42356974817175081 292 1/11 72/10000

For the latter I expected at least one or two months, but it came much earlier...

[robert44444uk]

Originally Posted by Thomas11

During the weekend I found the first VPS for E=196:

S 12371760550032321 196 100/8794 102/10000

And, even more exciting, the first E=292 Payam sequence:

S 42356974817175081 292 1/11 72/10000

For the latter I expected at least one or two months, but it came much earlier...

Very impressive, I would've thought a couple of months as well.

I'm travelling now for a week with no access to computers. But I started today on E130 again, as I think, with the huge pickup in sieve speed, this might become an optimal source for serious candidates for the higher prime counts.

[Thomas11]

Carlos,

I've run your latest results files through my script and got the following statistics:

100 452
101 426
102 310
103 232
104 182
105 106
106 67
107 44
108 24
109 18
110 8
111 7
112 10
113 3
114 2
115 0
116 1
117 1
total: 1893

I also checked your candidates for higher factors and found that the following are true E=82 sequences (verified with axn's payamx.exe):

R 67888682569393 66 == R 817935934571 82
R 133451144075469 66 == R 1607845109343 82
R 135673542891353 66 == R 1634620998691 82
R 141369385248117 66 == R 1703245605399 82
R 174048689082875 66 == R 2096972157625 82
R 189598044345983 66 == R 2284313787301 82
R 190560525020913 66 == R 2295909940011 82
R 240971051866039 66 == R 2903265685133 82
R 255566888563145 66 == R 3079119139315 82
R 258390921597515 66 == R 3113143633705 82
R 271292586795557 66 == R 3268585383079 82