OFFICIAL "SERVER PROBLEMS" THREAD

[retina]

This shows 6 exponents in the 77M range that are still pending (red), and 4 exponents are active (blue):

----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  Exponent Range    | Composite  | Status Unproven  |        Assigned         |        Available        |
   Start   Count  P |   F   LL-D | LL   LLERR NO-LL |  TF    P-1   LL   LL-D  |  TF    P-1   LL   LL-D  |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  77000000 55009  1 | 35812   600 18590     2     4 |                 4    34 |                   18557 |

Whereas this shows 3 assignments that are active:

Exponent	Work Type	Stage, % 		Expires	ETA 	Assigned	Last Update	Next Update	Estimated Completion	Userid
77864141	LL				6	3	2018-04-07	2018-04-07	2018-04-08	2018-04-11		Nick Thorpe
77959619	LL				6	9	2018-04-07	2018-04-07	2018-04-08	2018-04-17		PED
77998639	PRP		PRP, 43.50%		3	2018-01-28	2018-04-07	2018-04-08	2018-04-11		LookAS

What happened to the missing 3 exponents?

Also why don't PRP assignments have an expiry time?

[ixfd64]

I've been randomly seeing this message near the top of the home page today:

Warning: odbc_exec(): SQL error: [Microsoft][ODBC SQL Server Driver][SQL Server]Cannot insert the value NULL into column 'm_value', table 'primenet.dbo.t_gimps_milestones'; column does not allow nulls. UPDATE fails., SQL state 23000 in SQLExecDirect in C:\inetpub\www\default.php on line 24 Warning: odbc_fetch_array() expects parameter 1 to be resource, boolean given in C:\inetpub\www\default.php on line 25 Warning: odbc_free_result() expects parameter 1 to be resource, boolean given in C:\inetpub\www\default.php on line 28

However, the server functionality does not appear to be affected.

[retina]

Originally Posted by retina

This shows 6 exponents in the 77M range that are still pending (red), and 4 exponents are active (blue):

----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  Exponent Range    | Composite  | Status Unproven  |        Assigned         |        Available        |
   Start   Count  P |   F   LL-D | LL   LLERR NO-LL |  TF    P-1   LL   LL-D  |  TF    P-1   LL   LL-D  |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  77000000 55009  1 | 35812   600 18590     2     4 |                 4    34 |                   18557 |

Whereas this shows 3 assignments that are active:

Exponent	Work Type	Stage, % 		Expires	ETA 	Assigned	Last Update	Next Update	Estimated Completion	Userid
77864141	LL				6	3	2018-04-07	2018-04-07	2018-04-08	2018-04-11		Nick Thorpe
77959619	LL				6	9	2018-04-07	2018-04-07	2018-04-08	2018-04-17		PED
77998639	PRP		PRP, 43.50%		3	2018-01-28	2018-04-07	2018-04-08	2018-04-11		LookAS

What happened to the missing 3 exponents?

Also why don't PRP assignments have an expiry time?

Me is wondering if the data integrity is not at its finest? Can we be sure that no exponents have been missed in the past? I'm reasonably sure that the figures above are not normal, but I could be wrong.

[GP2]

As I write this, https://www.mersenne.org/primenet/ looks (in part) like this:

----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  Exponent Range    | Composite  | Status Unproven  |        Assigned         |        Available        |
   Start   Count  P |   F   LL-D | LL   LLERR NO-LL |  TF    P-1   LL   LL-D  |  TF    P-1   LL   LL-D  |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
  75000000 55307    | 35749   739 18819             |                 2    43 |                   18774 |
  76000000 54924    | 35603   613 18708             |                69    42 |                   18597 |
  77000000 55009  1 | 35812   601 18589     2     4 |                 4    35 |                   18555 |

And https://www.mersenne.org/assignments/ looks (in part) like this:

77864141	LL	LL, 58.40%	26	0	2018-04-07	2018-04-11	2018-04-12	2018-04-11	Nick Thorpe
77959619	LL	LL, 6.40%	26	7	2018-04-07	2018-04-11	2018-04-12	2018-04-18	PED
77231809	PRP	PRP, 6.00%		410	2017-12-17	2018-04-09	2018-04-10	2019-05-26	Kriesel
77998639	PRP	PRP, 44.90%		4	2018-01-28	2018-04-11	2018-04-12	2018-04-15	LookAS

Here's my interpretation:

The data show four pending assignments for 77M in both reports. In your case, the second report only showed three exponents, but perhaps the two reports were temporarily out of sync until one was updated.

Only the "Assigned" columns show actual assignments. The "Status Unproven" columns merely refer to exponents that haven't been verified with matching residues, which may or may not be currently assigned. The two LLERR could perhaps be "Suspect" first-time results which haven't yet been assigned for another LL test.

The report itself has some failings in that it doesn't take PRP into account properly as its own category.

In the 75M range,
the 2 under "Assigned LL" are actually PRP double-checks
the 43 under "Assigned LL-D" are LL double-checks, as expected

In the 76M range,
the 69 under "Assigned LL" are actually PRP double-checks (a few unfortunately also have a redundant LL result which will be wasted, due to glitches when the first PRP tests started getting done)
the 42 under "Assigned LL-D" are LL double-checks, as expected

In the 77M range, as we can see from the second report,
the 4 under "Assigned LL" are 2 pending first-time LL tests, one is a pending first-time PRP test, and one is a pending PRP self-double-check (or perhaps a hung exponent). Neither of those PRP test exponents has had an LL test done.

[chalsall]

Originally Posted by GP2

Here's my interpretation:

My interpretation is the Captain is asleep at the wheel.

This always works out so very well....

[Prime95]

Originally Posted by chalsall

My interpretation is the Captain is asleep at the wheel.

This always works out so very well....

Not far off the mark. While doing AVX512 development, I'm spending less time delving into PrimeNet mysteries.

[chalsall]

Originally Posted by Prime95

Not far off the mark. While doing AVX512 development, I'm spending less time delving into PrimeNet mysteries.

It's perfectly cool to resign. Become a consultant to the board.

Less work; more money.

Edit: Please don't take this the wrong way. None of us would be here if it wasn't for you.

[ATH]

Right now the status is:

  Exponent Range    | Composite  | Status Unproven  |        Assigned         |        Available        |
   Start   Count  P |   F   LL-D | LL   LLERR NO-LL |  TF    P-1   LL   LL-D  |  TF    P-1   LL   LL-D  |
  77000000 55009  1 | 35812   601 18590     2     3 |                 3    35 |                   18556 |

I ran a script counting exponent status from the text list of exponents status from 77M to 78M:
https://www.mersenne.org/report_expo...7001000&text=1

Results:

Primenet statistics
  77000000 55009  1 | 35812   601 18590     2     3
My count:
  77000000 55009  1 | 35813   601 18592     0     2

I got 18592 Unverified "LL" (158 PRP + 18434 LL) and 0 LLERR instead of 18590 LL and 2 LLERR.

That is because the 2 LL Suspect results has a successful PRP test which does not yet count in the statistics:
https://mersenne.org/M77732573
https://mersenne.org/M77879497


I got 35813 Factored instead of 35812, and 2 NO-LL instead of 3.

The 2 listed here 77959619 + 77998639:
https://www.mersenne.org/assignments...&exp1=1&extf=1


So that means that 1 of the 3 NO-LL might be factored but not counting for some reason.

[retina]

Originally Posted by GP2

And https://www.mersenne.org/assignments/ looks (in part) like this:

77864141	LL	LL, 58.40%	26	0	2018-04-07	2018-04-11	2018-04-12	2018-04-11	Nick Thorpe
77959619	LL	LL, 6.40%	26	7	2018-04-07	2018-04-11	2018-04-12	2018-04-18	PED
77231809	PRP	PRP, 6.00%		410	2017-12-17	2018-04-09	2018-04-10	2019-05-26	Kriesel
77998639	PRP	PRP, 44.90%		4	2018-01-28	2018-04-11	2018-04-12	2018-04-15	LookAS

77864141 has been completed so everything is reduced by one, but the server still has the same odd output.

I don't see that extra assignment for 77231809. This link only shows two assignments. 77959619 and 77998639.

In your output why is the ordering not correct? 77231809 appears out of order. Did the server output it in that order?

[ATH]

Originally Posted by retina

I don't see that extra assignment for 77231809. This link only shows two assignments. 77959619 and 77998639.

77231809 already has 1 PRP test completed so you have to remove the "Exclude double-check assignments" check mark to see the assignment:
https://www.mersenne.org/assignments...&exp1=1&extf=1

https://mersenne.org/M77231809

[retina]

Originally Posted by ATH

77231809 already has 1 PRP test completed so you have to remove the "Exclude double-check assignments" check mark to see the assignment:
https://www.mersenne.org/assignments...&exp1=1&extf=1

https://mersenne.org/M77231809

Okay, yes, you are right. So the problem there is that the server is counting that as an "Assigned LL" instead of "Assigned LL-D".

That now leaves us with the 2 missing tests, out of 5 total, that are listed as "Status Unproven".

What a mess that addition of PRP has become. Perhaps instead of identifying things by the type of test, it should be by the count of tests. So no more LL/LL-D columns, just have tested/verified columns.