[QUOTE=jinydu;141061]I see. So Ernst isn't personally doing a verification run.[/QUOTE]

Having spent 12 years coding Mlucas I can assure you the verification run is personally Ernst's as much as anyones :smile:

[quote=Kevin;140914]I'm aware of that. But an error that occurred before the last save file in some bizarre way that lead to a false positive would still lead to a false positive if you ran the last iterations again. As I already said, the optimism of George et. all in multiple statements are the best indications that there was no error.[/quote]
Bizarre is the word.
It is like being dealt a hand of 13 spades, only less likely.
I don't think anyone is capable of rigging this deck:smile:

[QUOTE=davieddy;141119]Bizarre is the word.
It is like being dealt a hand of 13 spades, only less likely.
I don't think anyone is capable of rigging this deck:smile:[/QUOTE]

Are you talking of playing Bridge? (I played bridge a few years ago for quite some time, but only "just for fun", and no "professional bidding" and such)

[quote=Andi47;141121]Are you talking of playing Bridge? (I played bridge a few years ago for quite some time, but only "just for fun", and no "professional bidding" and such)[/quote]
I used to play bridge quite seriously. I was once dealt a 4333 hand
with no card higher than a 7. Partner opened 2NT and there he played.
"Thankyou partner" was his disciplined response when I proudly laid
out my hand as dummy. It was a relatively high level duplicate match,
and a flat board.

[QUOTE=davieddy;141124]I used to play bridge quite seriously. I was once dealt a 4333 hand
with no card higher than a 7. Partner opened 2NT and there he played.
"Thankyou partner" was his disciplined response when I proudly laid
out my hand as dummy. It was a relatively high level duplicate match,
and a flat board.[/QUOTE]

Ouch! How many tricks did your partner make?

[QUOTE=davieddy;140929]0.0224 sec/iter means ten days for 40M iterations.
Taking this with the presumed inadequacy of a 2048K FFT,
I think we can anticipate an exponent > 40M.[/QUOTE]

Why do you presume 2048K FFT is inadequate, other than an independent estimate of the exponent?

Just for reference, in Prime95 version 24.14, 2048K starts at M34,560,000 and ends at M39,499,999. The next FFT level, 2.5M, starts at M39,500,000 and ends at M49,099,999. Of the 17 possible exponents, 6 are below 39.5M and 11 are above 39.5M. Ernst did advise against giving away the FFT size at one point, saying it would be too much of a hint...

Now, 40% done.
 

Now, 40% done.
Tony

[quote=jinydu;141136]Why do you presume 2048K FFT is inadequate, other than an independent estimate of the exponent?

Just for reference, in Prime95 version 24.14, 2048K starts at M34,560,000 and ends at M39,500,000.[/quote]
Because Ernst deemed 4096K necessary.

I think evrything about hiding and guessing the real exponent is kind of game or puzzle here. If primenet completely wanted to hide the exponent, they simply could list "1 prime unverified", but leave the prime candidate in assignment list as it was when updated the last time before checking in (and not move it to the cleared exponents list).

If moved to the cleared exponents list, the choice can be narrowed down to the list of all returned exponents in one hour (or maybe one day, if the update is not done immediately).
In that way the best way to hide is using a random residue. On the other hand a mathematic relation is more intersting for all puzzlers amung us.

When discovering M44, the fake residue was
32582657 68 L 0x663C8660956654__ 04-Sep-06 17:33 curtisc wd-102--04l

Did you notice that there is only one hexit (C) among 13 decimal numbers. that is quite unlikely (about 1%) but can happen.
That way i have chosen my guess.

It is possible that 2 people are verifying a prime number and they get the same wrong residue. if the chance of an error during a LLTest is 1%, the chance of a matching wrong residue is (1/10000)*2^-64.

[QUOTE]Prime, UNVERIFIED : 2[/QUOTE]

Check out this thread:
[url]http://www.mersenneforum.org/showthread.php?t=10616[/url]
Is it possible we just found M46 as well? Gotta admit, it sounds like it can't be, but hey it's there...

[QUOTE=biwema;141164]When discovering M44, the fake residue was
32582657 68 L 0x663C8660956654__ 04-Sep-06 17:33 curtisc wd-102--04l

Did you notice that there is only one hexit (C) among 13 decimal numbers. that is quite unlikely (about 1%) but can happen.
That way i have chosen my guess.[/QUOTE]

Wow, nice find! I'm surprised nobody else noticed that.

The residue of M42760397 (0x[B]A[/B]6090[B]C[/B]299[B]C[/B]0678__) has a strangely low proportion of hexadecimal digits. Does anyone else find it a bit... eye-catching?