mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > News

Reply
 
Thread Tools
Old 2016-01-13, 07:45   #166
axn
 
axn's Avatar
 
Jun 2003

491310 Posts
Default

Quote:
Originally Posted by M29 View Post
So, not the new prime itself, just that a new one exists (probably).
But even this has changed. Now even the fact of a new prime is hidden. The top producer's report used to show a success count for a new prime. Now... nothing. Really, they might as well have done the whole verification and press release silently instead of creating this zero-information thread.
axn is offline   Reply With Quote
Old 2016-01-13, 08:02   #167
Dubslow
Basketry That Evening!
 
Dubslow's Avatar
 
"Bunslow the Bold"
Jun 2011
40<A<43 -89<O<-88

11100001101012 Posts
Default

Quote:
Originally Posted by axn View Post
But even this has changed. Now even the fact of a new prime is hidden. The top producer's report used to show a success count for a new prime. Now... nothing. Really, they might as well have done the whole verification and press release silently instead of creating this zero-information thread.
But where's the fun in that?
Dubslow is offline   Reply With Quote
Old 2016-01-13, 08:11   #168
axn
 
axn's Avatar
 
Jun 2003

173 Posts
Default

Quote:
Originally Posted by Dubslow View Post
But where's the fun in that?
Where's the fun in this? Zero guesses. No one even bother to do "sleuthing".

Even before the thread was started, the prime had been verified. Independent verification happened in record time. Yet, here we are, none the wiser about the new prime. Fun? I don't think so.
axn is offline   Reply With Quote
Old 2016-01-13, 08:13   #169
Dubslow
Basketry That Evening!
 
Dubslow's Avatar
 
"Bunslow the Bold"
Jun 2011
40<A<43 -89<O<-88

3×29×83 Posts
Default

Quote:
Originally Posted by axn View Post
Where's the fun in this? Zero guesses. No one even bother to do "sleuthing".

Even before the thread was started, the prime had been verified. Independent verification happened in record time. Yet, here we are, none the wiser about the new prime. Fun? I don't think so.
And yet, were we to poll about having this thread or not, I suspect the response would be overwhelmingly affirmative.
Dubslow is offline   Reply With Quote
Old 2016-01-13, 08:46   #170
aketilander
 
aketilander's Avatar
 
"Åke Tilander"
Apr 2011
Sandviken, Sweden

2·283 Posts
Smile The verification runs

I get a Little confused about all verification runs going on and wonder which results have come in and which we are still waiting for. Below is what I have found in this thread:

Madpoo (Aaron Blosser)
Madpoo already verified it once before this thread started. Verified

We're at a point now where my 2nd run finished and still gets an "is prime" result
- As for the verification, I actually ran it a second time (saving the interim residues every 1M iterations, since I forgot the first time) Verified
- re-ran my penultimate save file (last 100K or so iterations) using a larger FFT of 5M and it did fine. Verified

ATH (Andreas Höglund)
ETA on 8 cores on 5960X ~ 8am UTC on tuesday the 12th.

ETA on Titan Black run ~ 10 am UTC on wednesday the 13th
Verified
- Titan Black ETA is now tuesday ~ 7 AM UTC, just before the CPU run.
- M( 7xxxxxxx )P, offset = 14508, n = 4000K, CUDALucas v2.05.1

airsquirrels (David S.)
Team AMD:
1100(OC) Mhz Fury X - Due ~ Tues 1/12/2016 10:27 PM EST
1000(UC) Mhz Fury X - Due ~ Weds 1/13/2016 4:50 AM EST

Team NVIDIA:
Titan Z (Air, One core, DP enabled) - Due ~ Tues 1/12/2016 8:59 AM EST
Titan (Water, DP enabled) - Due ~ Tues 1/12/2016 3:51 AM EST
Titan Black (Water, DP enabled) - Due ~ Tues 1/12/2016 2:41 AM EST

Team Intel:
Dual Xeon E5-2698 v3 + DDR4 (32 cores 1 worker) - Due Mon 1/11/2016 2:03 AM EST
Verified
- is prime" result, and AirSquirrels run with Prime95 was the same
- AirSquirrels finished his Intel + Prime95 run and we had matching residues along the way.

Right now AirSquirrels has a separate Intel + Prime95 run using a larger FFT, as well as the aforementioned GPU runs. Verified
- The Xeon run with a larger FFT and a separate CUDA run also came up positive.

Batalov + e.w.mayer Verified
ETA on Mlucas Run #1 is on wednesday the 13th, too.
ETA on Mlucas Run #2 is perhaps later but on a different FFT size and all residues, of course, can be matched between all runs.


Ernst run.

Last fiddled with by Batalov on 2016-01-13 at 14:27 Reason: (updated status)
aketilander is offline   Reply With Quote
Old 2016-01-13, 09:27   #171
bloodIce
 
bloodIce's Avatar
 
Feb 2010
Sweden

173 Posts
Default

Quote:
Originally Posted by Madpoo View Post
If that's a reference to my comment, look up hyperbole in the dictionary. Sorry, I thought it was obvious? Um... I wasn't *really* going to die or anything.

Or are you being funny and it totally whooshed over my head?
I did not intend any insult to you or anyone else. I just pointed apparently in an inappropriate way, that most of the guys knowing the exponent are quite secretive about it. Verification is nice to have, but the whole thread looks like: "Some of us know something secret, but we will not share it with you." May be you should have contacted the verifiers in private in advance and out that there is a prime coming in a press release in a day or two. That is fair. Now it looks like a mess. On top, some of you knowing the exponent are pointing how wrong are the guesses :-). That is quite cynical.

P.S. Personally, I think it is quite disrespectful to the community to emphasize how well you hidden the find from your helpers. Do not forget that the prime was supposedly found by anonymous user in attempt to help the project. How would the rest of the helpers feel about that?

Last fiddled with by bloodIce on 2016-01-13 at 09:59 Reason: P.S.
bloodIce is offline   Reply With Quote
Old 2016-01-13, 10:06   #172
aketilander
 
aketilander's Avatar
 
"Åke Tilander"
Apr 2011
Sandviken, Sweden

2×283 Posts
Smile

Quote:
Originally Posted by bloodIce View Post
That is quite cynical.
Actually I Think it is very good that we have a thorough verification process Before announcing the exponent. Lets say we annonced it first and then it later turned out to be a false positive, the quality of the whole GIMPS-Project would be questioned. We don't need something like a "Skylark bug"-discussion which put the results from the Project into question.

One thing I have been thinking about is if there is a way to build a version/mode of prime95 to be (even more) rock-solid, to be used (only) for verification issues like this. Maybe with built in checking and double-checking on every iteration (with a lot of cores this would not slow down the process) and error checking on every iteration (with a lot of cores, this again, would not slow down the DC much)? Maybe there are other ways also to build in more error checking in the assembler code or make the program (even more) robust? It may be useful also for the real big LL runs.
aketilander is offline   Reply With Quote
Old 2016-01-13, 10:08   #173
lycorn
 
lycorn's Avatar
 
Sep 2002
Oeiras, Portugal

1,451 Posts
Default

@bloodice:
Honestly, I can´t see any disrespect in doing that. It has been explained ten times over why the potential primes are hidden until the verifications deemed adequate are run and the press release is ready. If some mechanism was put in place to prevent the early leaking of the news, so much the better it is good and works. Feeling "disrespected" due to that is in my opinion a bit childish. (Note that I am in the very same situation as yourself, I do not have any clue about the number in question).

Last fiddled with by lycorn on 2016-01-13 at 10:08
lycorn is offline   Reply With Quote
Old 2016-01-13, 10:32   #174
ewmayer
2ω=0
 
ewmayer's Avatar
 
Sep 2002
República de California

3×53×31 Posts
Default

Quote:
Originally Posted by aketilander View Post
I get a Little confused about all verification runs going on and wonder which results have come in and which we are still waiting for. Below is what I have found in this thread:
[snip]

Batalov
ETA on Mlucas Run #1 is on wednesday the 13th, too.
ETA on Mlucas Run #2 is perhaps later but on a different FFT size and all residues, of course, can be matched between all runs.


ewmayer
Ernst run.
2 small notes:

o AFAIK the Prime95 DCs are not considered 'independent verify' because even they they use different shifts for the initial 'rotated residue', they use the same code. I'll let the DCers in question comment on whether any of the runs used an FFT length different than 4096K.

o I'm not doing a verify run; Serge is using my code (specifically a Haswell-style AVX2/FMA3 binary thereof) to do one via his Amazon Web Services cloud account. As you note he is doing 2 separate DCs, one @4096K and one @4608K. The smaller-FFT one should finish in around 2-3 hours - original ETA was 3:50am PST, but he lost an hour yesterday evening when his run got borked due to Free Market Capitalism - and the larger-FFT run is proceeding ~20% slower. Due to the FFT-size-independence of the Mlucas bytewise savefile format, not only can we compare Res64 values at various checkpoints, we can in fact directly compare the entire residues between the various 10Miter-multiple persistent savefiles deposited by the program, and between the final 100Kiter-multiple checkpoint files, simply using Linux 'diff'. (And Serge, that means you only need one copy of each of the aforementioned savefiles in your coming public-share upload, along with a copy of both .stat files, for reference purposes.)

Off to bed - assuming the result is confirmed a few hours from now, congrats to GIMPS, George, and the lucky discoverer(s)!
ewmayer is offline   Reply With Quote
Old 2016-01-13, 10:56   #175
aketilander
 
aketilander's Avatar
 
"Åke Tilander"
Apr 2011
Sandviken, Sweden

2×283 Posts
Smile

Thank you for the clarification Ernst!

Did anyone set up the old Glucas for a verification run? There was some talk about this before in this thread.
aketilander is offline   Reply With Quote
Old 2016-01-13, 11:04   #176
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

41×229 Posts
Default

Quote:
Originally Posted by ewmayer View Post
...(And Serge, that means you only need one copy of each of the aforementioned savefiles in your coming public-share upload, along with a copy of both .stat files, for reference purposes.)
Yep, I do have them all. They will be available @ dropbox.
Batalov is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Are Bitcoins Prime Related a1call Miscellaneous Math 26 2021-03-18 14:18
Merry Christmas and a prime! (M50 related) Prime95 News 505 2020-01-18 01:03
Oops i did it again. (Prime found) ltd Prime Sierpinski Project 21 2006-01-04 14:50
Another new prime (M42-Related) Uncwilly News 132 2005-05-10 19:47
some prime-related trick questions ixfd64 Puzzles 2 2003-09-23 12:53

All times are UTC. The time now is 10:04.

Thu Apr 15 10:04:02 UTC 2021 up 7 days, 4:44, 0 users, load averages: 2.42, 1.70, 1.32

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.