OK I thought that the amount of data in the future would be too large and I decided to do the sieving in C.

Using the available scanning data, I could only go up to 2161 : n1=46 and n2=2114.
The first hicup comes at 1511 so there is no point in dealing with primes up to 1 million so far. I need way more scanning data.

So far the currently known sieve is:

11 3  7
19 4  14
29 5  23
31 12  18
41 6  34
59 25  33
61 17  43
71 8  62
79 29  49
89 9  79
101 22  78
109 10  98
131 11  119
139 63  75
149 40  108
151 27  123
179 74  104
181 13  167
191 88  102
199 61  137
211 32  178
229 81  147
239 15  223
241 51  189
251 117  133
269 71  197
271 16  254
281 37  243
311 58  252
331 116  214
349 143  205
359 105  253
379 19  359
389 151  237
401 111  289
409 129  279
419 20  398
421 110  310
431 90  340
439 69  369
449 165  283
461 21  439
479 228  250
491 73  417
499 224  274
509 121  387
521 99  421
541 172  368
569 232  336
571 273  297
599 24  574
601 136  464
619 242  376
631 109  521
641 278  362
659 200  458
661 57  603
691 221  469
701 26  674
709 170  538
719 329  389
739 118  620
751 210  540
761 91  669
769 338  430
809 342  466
811 28  782
821 212  608
829 95  733
839 341  497
859 276  582
881 326  554
911 67  843
919 316  602
929 30  898
941 227  713
971 173  797
991 31  959
1009 382  626
1019 493  525
1021 457  563
1031 106  924
1039 286  752
1049 325  723
1051 72  978
1061 459  601
1069 275  793
1091 211  879
1109 406  702
1129 327  801
1151 558  592
1171 113  1057
1181 533  647
1201 77  1123
1229 484  744
1231 269  961
1249 404  844
1259 35  1223
1279 599  679
1289 156  1132
1291 627  663
1301 267  1033
1319 399  919
1321 452  868
1361 82  1278
1381 609  771
1399 239  1159
1409 124  1284
1429 546  882
1439 700  738
1451 282  1168
1459 166  1292
1471 477  993
1481 38  1442
1489 680  808
1499 208  1290
1531 87  1443
1549 529  1019
1559 39  1519
1571 567  1003
1579 682  896
1601 605  995
1609 635  973
1619 764  854
1621 175  1445
1669 135  1533
1699 229  1469
1709 600  1108
1721 41  1679
1759 858  900
1801 426  1374
1811 185  1625
1889 823  1065
1901 97  1803
1951 146  1804
1979 44  1934
1999 949  1049
2011 735  1275
2069 45  2023
2089 611  1477
2099 665  1433
2129 792  1336
2161 46  2114

Of course, moduli 4 and 5 are treated separately. With now 143 known results, this program has decupled the size of the available sieve so it's now easier to finely check candidate moduli.

And as usual the source code is there : create_sieve.tgz.

.