New milestone

[tha]

From the first 10M digits Mersenne prime candidate upwards a range of 212.100 has been tested at least once.

All exponents between 33.219.127 and 33.431.227 have either been factored or Lucas-Lehmer tested once. Some tests have error codes and are currently server assigned double checked ahead of the regular double check range. Other exponents in this range have been successfully double checked already.

My main machine is currently crunching on the next two lowest 10M digits exponents. The results will be checked in at the end of this month.

[jinydu]

Just moments ago, I was assigned my very first double-checking assignment larger than 19M. I'm not sure whether to consider that good news or bad news (GIMPS is moving forwards, but the tests are taking ever longer to complete)...

[cheesehead]

Originally Posted by jinydu

I'm not sure whether to consider that good news or bad news (GIMPS is moving forwards, but the tests are taking ever longer to complete)...

Well, one of those is good news and the other is a natural consequence. :-)

[tha]

Originally Posted by tha

From the first 10M digits Mersenne prime candidate upwards ...

all exponents up to (therefore not including) 33,490,007 have been checked at least once. That means a range of 270,880 has been cleared. I am currently working on the next two exponents.

[S485122]

Originally Posted by tha

all exponents up to (therefore not including) 33,490,007 have been checked at least once.

I do not understand ? You mean the 32M numbers I am testing have changed in doublechecks ?

As far as I can see GIMPS mentions some 23 numbers below 20,4M as untested and PrimeNet lists some 14M numbers as untested. (PrimeNet is up to date but counts only those tests it assigned, the data from GIMPS is a few days old.)

Jacob

[davieddy]

He's talking about exponents with 10 million digits

[tha]

All exponents between 33.219.127 and 33.500.837 have been tested at least once, spanning a range of 281.710. A few percent has been double checked and a few exponents have results with error codes other than zero without a matching double check.

[petrw1]

Then why do I still see these 18 first-time LL assignments on the status file?

33219127 69 24232000 223.6 16.8 59.8 08-Apr-07 15:45 29-Aug-06 05:54 S424555 GeidiPrime
33230963 69 355.6 49.3 78.3 09-Mar-07 04:07 19-Apr-06 06:38 S16447 shum
33249343 69 18517696 671.5 -3.5 56.5 14-Mar-07 08:24 07-Jun-05 07:06 S16447 aav
33271549 69 7178426 15.5 14.8 72.8 25-Mar-07 14:54 25-Mar-07 07:45 S102309 C60657715
33279397 69 7834768 18.8 10.2 69.2 22-Mar-07 01:14 wmjmoeller conrac01
33288743 69 16874981 6.2 11.8 68.8 03-Apr-07 15:23 wherbert Sonya2400
33289229 69 10931776 67.0 19.9 60.9 09-Apr-07 17:37 01-Feb-07 20:38 Malloe AS2G
33322867 69 26048030 758.1 63.7 85.7 07-Apr-07 12:01 12-Mar-05 18:11 hopey C8AC1010B
33386203 69 6449568 1561 369.3 65.3 01-Apr-07 04:35 31-Dec-02 03:53 GTS GTS
33411179 69 610094 25.0 79.8 56.8 05-Apr-07 16:47 15-Mar-07 20:08 S19490 C92281B84
33431227 69 80.7 37.5 60.5 09-Apr-07 07:00 19-Jan-07 04:59 kmr tiger1
33460759 69 80.6 106.4 97.4 28-Mar-07 05:43 19-Jan-07 05:19 S16447 peg
33469109 69 1337600 80.6 12.9 61.9 14-Mar-07 18:38 19-Jan-07 05:13 RTR007 QV-1
33475369 69 13705728 1285 136.9 62.9 05-Apr-07 17:05 02-Oct-03 14:46 simba1999 jochen
33484673 69 14892417 510.2 110.3 54.3 03-Apr-07 02:23 15-Nov-05 16:00 Obsidian D4100pc
33489727 69 6395648 80.6 7.3 34.3 14-Mar-07 04:26 19-Jan-07 05:13 RTR007 QV-2
33490007 70 1043201 251.0 51.7 59.7 08-Apr-07 13:59 01-Aug-06 20:23 darkshikari Athlon64
33498719 69 7151976 570.7 -41.5 18.5 31-Jan-07 07:32 16-Sep-05 03:12 bjtags bjtags

[tha]

Originally Posted by petrw1

Then why do I still see these 18 first-time LL assignments on the status file?

I did not check them but most or all of them will have been from the batches that were reassigned due to error codes <> 0 returned with the results. So there is a result known, but the result is not considered reliable enough to pass for a first time LL test.

[petrw1]

On a related point ...
I just noticed that on the PrimeNet Individual Accounts Page where it defines the various test it says:

"Double-check assignments are identical to Lucas-Lehmer assignments, however a (D) also appears to the right of the exponent. These comparatively rare tests recompute earlier residues of questionable reliability."

... comparatively rare tests ... of questionable reliability ...

Does this imply that NOT ALL LL tests are double checked. In fact this statement suggests to me that MOST ARE NOT double checked.

On the other hand the: http://www.mersenne.org/math.htm page states:

"To verify that a first-time Lucas-Lehmer primality test was performed without error, GIMPS runs the primality test a second time."

This statement suggest to me that ALL are double checked?

As well when I look at the numbers in the "Double Check Avail" column it looks just as high as the numbers in the "LL Avail" column, again suggesting ALL exponents are double checked?

Do you what is the correct answer?

[cheesehead]

All LL tests are double-checked, eventually.

Originally Posted by petrw1

"Double-check assignments are identical to Lucas-Lehmer assignments, however a (D) also appears to the right of the exponent. These comparatively rare tests recompute earlier residues of questionable reliability."

... comparatively rare tests ... of questionable reliability ...

It is rare, however, to find an ace programmer who is also an ace technical writer who obsessively reviews all web page wording every time the project is tweaked.

Does this imply < snip >

No, it just imp^H^H^Hreveals that this is a volunteer-run project. :-}

Thank you for bring that wording to our collective attention, petrw1 !!