And now for some TF results...

lycorn · 2019-02-22, 07:55

In the 06:00 AM UTC "Results" report, a user identified as Samantha reported over 2600 TF results. Each one is worth 1088 ghz-days and consists of factoring numbers with exponents in the lower 900 M range from 1 to 82 bits. This amounts to a credit of over 3 million GHz-days, as per the Top Producers TF report.
Not a single factor was found. Although obviously possible, it is, to say the least, a very unusual situation, as the numbers had (at least the ones I checked) been previously tried to only 71 bits. 2600+ tests covering 11 bit levels without a single factor? Over 3 million Ghz-days of work reported at once? That makes me go hmmm...

ramgeis · 2019-02-22, 16:40

Quote:

Originally Posted by lycorn

In the 06:00 AM UTC "Results" report, a user identified as Samantha reported over 2600 TF results. Each one is worth 1088 ghz-days and consists of factoring numbers with exponents in the lower 900 M range from 1 to 82 bits. This amounts to a credit of over 3 million GHz-days, as per the Top Producers TF report.
Not a single factor was found. Although obviously possible, it is, to say the least, a very unusual situation, as the numbers had (at least the ones I checked) been previously tried to only 71 bits. 2600+ tests covering 11 bit levels without a single factor? Over 3 million Ghz-days of work reported at once? That makes me go hmmm...

This also raised my eyebrows... So I took the first 100 exponents in question and TFed from 71 to 72.
I found already two factors for https://www.mersenne.org/M900141871 and https://www.mersenne.org/M900142097
I propose to delete all the results submitted by that user as they obviously are bogus.

lycorn · 2019-02-22, 17:00

Seconded.

kriesel · 2019-02-22, 21:14

Quote:

Originally Posted by lycorn

In the 06:00 AM UTC "Results" report, a user identified as Samantha reported over 2600 TF results. Each one is worth 1088 ghz-days and consists of factoring numbers with exponents in the lower 900 M range from 1 to 82 bits. This amounts to a credit of over 3 million GHz-days, as per the Top Producers TF report.

That's enough credit to register as third in the TF top producers list, briefly, until the effect of what appears to be falsified results is (was?) removed.

LaurV · 2019-02-23, 02:23

Thirded. (is this even a word?)
All results should be deleted and the used given a negative credit, like -200 or so, so he has to do some work to be able to "start clean". It was done few times in the past for other cheaters.

Edit: Maybe somebody with more "spidering" abilities can make a list of Samantha's work? I will pick few of them for few bitlevels when I get rid of the current gpu72 work.

Madpoo · 2019-02-23, 04:45

User was created only a day ago with a clearly made up email address (I've lobbied before that we should verify email addresses by emailing a verification link...)

So, someone creates a new account with the sole purpose of being a jerk. Meh... it's the Internet. This kind of stuff happens, we just need to deal with it, and it's weird that someone would bother. What's the point?

Anyway, George pointed it out earlier today so I'm here to get more details...

5520 TF results, all of them "no factor found". It'd be astounding to do that many and not find any factors, ya know? And making a point of specifying "1 to 82 bits", because all of them had been TF'd to at least some bit level in the 60's, haven't they?

Plus, with a line like this, doesn't this signify that the mfaktc build only goes up to 76 bits?

Code:

no factor for M900200747 from 2^1 to 2^82 [mfaktc 0.20 barrett76_mul32_gs]

I hope this isn't the same little hax0r kid who copied/pasted the same result over and over, thousands of times. Hopefully not, but I guess now I'll have to check the logs and see, just in case.

Besides these recent cases of obvious malfeasance, I was also dealing with a user who was doing some very aggressive crawling of the exponent reports, tens or even hundreds of thousands of requests per day. No idea why or what kind of data they were after. I emailed and said if you need some data, just ask. Much easier than crawling millions of exponent report pages one by one.

Anyway... the thrills and chills of server admin work.

EDIT: "Samantha" appears to be in Stockholm... Meanwhile, I'm planning to just set each of those 5520 exponents back to the TF level they were at before and I'll just nuke all those bogus results.

De Wandelaar · 2019-02-23, 06:42

En today user emily … with 33 millions GHz/day
TF in the range 612 millions and 614 millions, 1 to 84 bits and systematically 'no factor found'.

Chuck · 2019-02-23, 07:04

Quote:

Originally Posted by De Wandelaar

En today user emily … with 33 millions GHz/day
TF in the range 612 millions and 614 millions, 1 to 84 bits and systematically 'no factor found'.

And of course the "emily" user link leads to a porn site.

Madpoo · 2019-02-23, 08:13

Quote:

Originally Posted by Chuck

And of course the "emily" user link leads to a porn site.

Just toasted those results too.... this time from Moldova. If it's the same user as "samantha" (seems highly likely), they're hopping around on a VPN. Or they merely travel between Sweden and Moldova just for fun. Yeah, that's probably it.

retina · 2019-02-23, 08:28

Quote:

Originally Posted by Madpoo

User was created only a day ago with a clearly made up email address (I've lobbied before that we should verify email addresses by emailing a verification link...)

Verifying email addresses won't fix what you want here. There are many temporary email address domains that anyone can use to create such accounts.

Verifying email addresses is to stop someone using another person's address by inputting their email address (perhaps by mistake, or malice, or whatever).

So okay, go ahead an verify email addresses, but it wont solve this particular problem.

Madpoo · 2019-02-23, 08:51

Quote:

Originally Posted by retina

Verifying email addresses won't fix what you want here. There are many temporary email address domains that anyone can use to create such accounts.

Verifying email addresses is to stop someone using another person's address by inputting their email address (perhaps by mistake, or malice, or whatever).

So okay, go ahead an verify email addresses, but it wont solve this particular problem.

True, but it's more about slowing them down. Basically making it harder to do stupid things so they find easier targets to waste time on.

Right now, whoever this is, can just keep creating sock puppet accounts and we can play whack-a-mole. But if they had to keep ginning up new email addresses to play the game, it takes more of their time.

Plus it gives us a bit more of an audit trail... if they're using the same freebie email service repeatedly it could make them easier to spot in advance, if it's strange enough.

Both of the IP addresses used show up in various abuse databases, making it seem more likely these are VPNs from some service that doesn't care what their users do. Those services can be useful sometimes... slap some code on a front-end or on the web server itself that looks up each IP and blocks any with a bad enough reputation. That would have stopped these, perhaps. But those services aren't free at the scale of traffic mersenne.org gets. And based on past experience, the risk of false positives is > 0.

Anyway, we've got a couple other ideas we're spitballing... any suggestions from the hoi polloi?

2019-02-22, 07:55	#1
lycorn Sep 2002 Oeiras, Portugal 2³·181 Posts	And now for some TF results... In the 06:00 AM UTC "Results" report, a user identified as Samantha reported over 2600 TF results. Each one is worth 1088 ghz-days and consists of factoring numbers with exponents in the lower 900 M range from 1 to 82 bits. This amounts to a credit of over 3 million GHz-days, as per the Top Producers TF report. Not a single factor was found. Although obviously possible, it is, to say the least, a very unusual situation, as the numbers had (at least the ones I checked) been previously tried to only 71 bits. 2600+ tests covering 11 bit levels without a single factor? Over 3 million Ghz-days of work reported at once? That makes me go hmmm... Last fiddled with by lycorn on 2019-02-22 at 07:57

2019-02-22, 17:00	#3
lycorn Sep 2002 Oeiras, Portugal 2³×181 Posts	Seconded. Last fiddled with by lycorn on 2019-02-22 at 17:01

2019-02-23, 02:23	#5
LaurV Romulan Interpreter Jun 2011 Thailand 22070₈ Posts	Thirded. (is this even a word?) All results should be deleted and the used given a negative credit, like -200 or so, so he has to do some work to be able to "start clean". It was done few times in the past for other cheaters. Edit: Maybe somebody with more "spidering" abilities can make a list of Samantha's work? I will pick few of them for few bitlevels when I get rid of the current gpu72 work. Last fiddled with by LaurV on 2019-02-23 at 02:27

2019-02-23, 04:45	#6
Madpoo Serpentine Vermin Jar Jul 2014 CDA₁₆ Posts	User was created only a day ago with a clearly made up email address (I've lobbied before that we should verify email addresses by emailing a verification link...) So, someone creates a new account with the sole purpose of being a jerk. Meh... it's the Internet. This kind of stuff happens, we just need to deal with it, and it's weird that someone would bother. What's the point? Anyway, George pointed it out earlier today so I'm here to get more details... 5520 TF results, all of them "no factor found". It'd be astounding to do that many and not find any factors, ya know? And making a point of specifying "1 to 82 bits", because all of them had been TF'd to at least some bit level in the 60's, haven't they? Plus, with a line like this, doesn't this signify that the mfaktc build only goes up to 76 bits? Code: no factor for M900200747 from 2^1 to 2^82 [mfaktc 0.20 barrett76_mul32_gs] I hope this isn't the same little hax0r kid who copied/pasted the same result over and over, thousands of times. Hopefully not, but I guess now I'll have to check the logs and see, just in case. Besides these recent cases of obvious malfeasance, I was also dealing with a user who was doing some very aggressive crawling of the exponent reports, tens or even hundreds of thousands of requests per day. No idea why or what kind of data they were after. I emailed and said if you need some data, just ask. Much easier than crawling millions of exponent report pages one by one. Anyway... the thrills and chills of server admin work. EDIT: "Samantha" appears to be in Stockholm... Meanwhile, I'm planning to just set each of those 5520 exponents back to the TF level they were at before and I'll just nuke all those bogus results. Last fiddled with by Madpoo on 2019-02-23 at 05:34

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Results	ET_	Operazione Doppi Mersennes	627	2021-01-31 18:06
Where have all the TF results gone?...	lycorn	PrimeNet	22	2017-10-02 02:40
PGS Results	danaj	Prime Gap Searches	0	2017-08-14 18:35
CPU Results last 24 hrs	Unregistered	Information & Answers	3	2010-07-26 00:49
0x results...	Mike	PrimeNet	11	2004-05-23 12:55

2019-02-23, 06:42	#7
De Wandelaar "Yves" Jul 2017 Belgium 3×17 Posts	En today user emily … with 33 millions GHz/day TF in the range 612 millions and 614 millions, 1 to 84 bits and systematically 'no factor found'.