Well, do you?

EDIT: Follow this link to get to the real introduction to this project: https://www.primegrid.com/forum_thread.php?id=8422&nowrap=true#125716
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

The question is, do you?

Also, the numerical representation of the date today DD/MM (29/01) is the same set of digits, in a different order, as the year YYYY (2019)
____________
My Primes
Badge Score: 4*2 + 6*2 + 7*3 + 8*10 + 11*3 + 12*1 = 166

Oh yes.

Yep, I'm pretty sure I just downloaded the next SR5 prime. Still haven't finished, though, or even started the task for that matter.

Sure : I found a PPSE before starting TdP!

So did I - first T5K prime for 6.5 years for me!
____________

Last year I got my first ever prime in TDP on a whim and I haven't achieved any other prime since, so, yeah, I'm guessing February is the best month for primes :D

Given that we're three mega-primes behind where we were last year, maybe we aren't so lucky. Unless we find three more mega primes in the next two days.

I am also enjoying how such a short and vague post has sparked an emphatic reaction.

Also, a misquoted Dirty Harry approves of this thread.
____________
My Primes
Badge Score: 4*2 + 6*2 + 7*3 + 8*10 + 11*3 + 12*1 = 166

Well, do you?

It's definitely GFN, it's in the GFN graphs in the server info - task status.

I'ld say my best guess is either a GFN-23 or starting from a big enough "b" – a GFN-22-100M-prime search (or just a GFN-22-WR, but you would still need to start from a solid point, not just random "wherever the WR is at right now"). Like we have GFN-17Low and GFN-17Mega.

"Do You Feel Lucky?" subproject is also mentioned on the editing your settings page in context that it doesn't have CPU apps.

Luck does not exist.

Luck does not exist.

Then Robish doesn't exist ;-)

Luck does not exist.

Then Robish doesn't exist ;-)

Hahahaha, yes I'm here, I'm real, I promise :))))

Well, do you?

Well, do you?

Since XGFN jobs pending 1 still stand: I am sure that 321 Prime Search is found prime. :)
And since I didnot run it , I cannot be that lucky person :)

Well, do you?

Since XGFN jobs pending 1 still stand: I am sure that 321 Prime Search is found prime. :)
And since I didnot run it , I cannot be that lucky person :)

It was 0 when I looked after you first wrote.

If anyone still isn't sure what kind of thing this thread is about, I advise you to look closely at the list of the subprojects on the front page.

Do You Feel Lucky?

In 2009, PrimeGrid started running GFN-15 and GFN-16 on PRPnet using Yves Gallot's Genefer program.

In 2010, GFN-18 and GFN-19 were also started on PRPNet. At that time, only one GFN-18 prime was known, and no GFN-19 primes were known. In February and March of 2011, PrimeGrid found the second and third known GFN-18 primes. Then, in October, we found the 4th known GFN-18 prime: 361658²⁶²¹⁴⁴+1.

361658²⁶²¹⁴⁴+1 is special: It's the first prime that we found using the new GeneferCUDA GPU program. The GPU version of Genefer was about 40 times faster than the CPU version, and represented a tremendous increase in computing power.

Right after that discovery, John challenged us to use GeneferCUDA to find a GFN-19. Nobody knew if a GFN-19 even existed. Certainly, none had ever been discovered.

"Why not?" said I. 4 days and just 16 tasks later, I discovered the world's first GFN-19 prime.

At that point I had been crunching at PrimeGrid for a bit less than two years. Somewhere during those two years I learned about GIMPS. Their Mersenne prime search had found all the largest primes. The primes found at PrimeGrid, as big as they were, were still much smaller than the Mersenne primes.

Seeing a chance to perhaps find primes as large as those found by GIMPS, I suggested that if we could make GeneferCUDA work on our BOINC server, where it would be more heavily utilized, we might start searching the GFN-22 range. All but the smallest potential GFN-22 primes would be larger than the then world record Mersenne prime, which was 12.9 million digits. Considering how fast GeneferCUDA was, we had a real shot at finding a world record prime. It was still a long shot, but there was a chance.

I offered to port Genefer and GeneferCUDA to be native BOINC programs, an idea to which the PrimeGrid admins were amenable. So I ported the programs and they made the necessary changes to PrimeGrid's servers. We started running GFN-18 on BOINC in January 2012, and started the world record prime search with GFN-22 in February 2012.

A year later, in February 2013, GFN-22 was searching numbers of about 16.5 million digits when GIMPS discovered a new world record Mersenne prime of 17.4 million digits.

Three years later they found a 22.3 million digit Mersenne prime.

Two years after that they found a 23.2 million digit Mersenne prime.

Eleven months later, in December of last year, they found a 24.8 million digit prime, which is the current largest known prime number.

GFN-22 is currently searching 21.6 million digit numbers.

It would appear that the opportunity to find a world record prime with GFN-22 has come and gone. Even though we have an excellent GPU program, Mersenne has the advantage of being inherently easier to search, and their numbers grow in size faster because their search increases n (size is proportional to n) whereas a GFN search increases b (size is proportional only to log(b)). In addition, they only need to search prime values of n whereas we need to search all even values of b. Even though we have a faster program, their search still progresses faster overall. We're behind, and they're pulling away.

After each of those 4 Mersenne discoveries, there were people here -- smart people -- who felt we should either increase b to be once again searching for a world record, or start an n=23 search, which would also be a world record search. The math, however, just doesn't make that a good bet. When we started GFN-22 it wasn't long before it was significantly larger than the current world record. We had some leeway. Starting a new world record search with larger numbers would have an even smaller chance of success. My uncle used to say, "Lotto is a tax on the mathematically challenged." I feel the same way about boosting GFN higher to be above GIMPS. It's not a good bet. They're more likely to find another prime before we find one of that size.

So after each Mersenne prime, we've thought about making another world record search, and each time, we decided not to do it.

Until now.

Let me start by saying this is a real long shot. It borders on stupidity. But a lot of people want to do it, and it's going to be fun.

We're starting up a new project called "Do You Feel Lucky?" It's going to be a GFN-22 search with b starting at 846,398. That's the first b not removed by sieving where b^4194304+1 is greater than the current world record.

Like GFN-17, we'll now have two GFN-22 searches running concurrently. Unlike GFN-17, however, this new search will not be a contiguous search. If a new Mersenne prime is found, we will increase b again so that we continue to search for a world record. GFN-22 is sieved to b=100M, so we can continue doing this search up to about 33.5 million digits. If we have to search larger numbers, we'll either need to sieve GFN-22 above 100M, or sieve GFN-23.

"Lucky?" tasks will be slower than normal GFN-22 tasks because neither the OCL nor OCL4 transforms can be used. All GPUs should use OCL5.

CPU tasks will not be available. These tasks are beyond the range of the 64 bit CPU transforms, so the 80 bit x87 transform must be used. Not only is this about 10 times slower, but it can't use multi-threading like the AVX and FMA3 CPU transforms can. CPUs would take forever to run these tasks.

This project is a prime-finding project, so it will be available for TdP, or at least most of it. I'm not sure exactly when we'll turn it on.

The internal name is "genefer_extreme". That's what you would use in app_config.xml or app_info.xml. Should it ever be necessary, the next names would be genefer_ultimate, then genefer_ludicrous, then genefer_plaid.
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

The internal name is "genefer_extreme".

I'd like to offer special thanks to two people.

The first is, of course, Yves Gallot, without whom none of this would be possible. Not only did Yves write the original Genefer program, but his subsequent Genefer OCL and improved CPU programs made it possible for us to expand the Genefer searches in ways that were previously impossible.

The second person is Robish, who is simply the luckiest GFN searcher on the planet. If anyone is going to find a GFN-22, it's going to be him. Yes, this is actually the reason we said yes this time. We're going to need a ridiculous amount of luck to find a world record prime, and Robish seems to have the market cornered.

No pressure, Rob. :)
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

I'd like to offer special thanks to three people.



The second person is Robish, who is simply the luckiest GFN searcher on the planet. If anyone is going to find a GFN-22, it's going to be him. Yes, this is actually the reason we said yes this time. We're going to need a ridiculous amount of luck to find a world record prime, and Robish seems to have the market cornered.

No pressure, Rob. :)

I'd like to offer special thanks to three people.



The second person is Robish, who is simply the luckiest GFN searcher on the planet. If anyone is going to find a GFN-22, it's going to be him. Yes, this is actually the reason we said yes this time. We're going to need a ridiculous amount of luck to find a world record prime, and Robish seems to have the market cornered.

No pressure, Rob. :)

Rob, what are you waiting :)))
Go go go :)
Lets find that prime!

Wow, just wow! I dont feel lucky (although I'm somewhat mathematically challenged), but my brain is telling to get a faster GPU asap...


On a side note, Michael, do you you mind if I quote your uncle's wise words elsewhere?

I'm in! Absolutely!
:)

Is it unlikely?..... Yes. Is it impossible? Eh no. Mike proved that after only 16 units on gfn19.

If you're on the fence, just imagine the thrill of getting one!
Lotto ticket?, yes, but for the PG community, massive cudos and that's why we're all here I think. To contribute to the community that is PG. And hopefully break a few records on the way :)

I look forward to congratulating anyone who finds one :)

This is going to be fun.
____________
My lucky number 1059094^1048576+1

I'd like to offer special thanks to three people.



The second person is Robish, who is simply the luckiest GFN searcher on the planet. If anyone is going to find a GFN-22, it's going to be him. Yes, this is actually the reason we said yes this time. We're going to need a ridiculous amount of luck to find a world record prime, and Robish seems to have the market cornered.

No pressure, Rob. :)

Rob, what are you waiting :)))
Go go go :)
Lets find that prime!

I found the GFN-19 in 16 tasks. Can he find the GFN-22 in 15 or fewer tasks?

Do You Feel Lucky?

This project is a prime-finding project, so it will be available for TdP, or at least most of it. I'm not sure exactly when we'll turn it on.

The internal name is "genefer_extreme". That's what you would use in app_config.xml or app_info.xml. Should it ever be necessary, the next names would be genefer_ultimate, then genefer_ludicrous, then genefer_plaid.

Will it be added to the existing GFN badge or will there be a new badge created?

Do You Feel Lucky?

This project is a prime-finding project, so it will be available for TdP, or at least most of it. I'm not sure exactly when we'll turn it on.

The internal name is "genefer_extreme". That's what you would use in app_config.xml or app_info.xml. Should it ever be necessary, the next names would be genefer_ultimate, then genefer_ludicrous, then genefer_plaid.

Will it be added to the existing GFN badge or will there be a new badge created?

It will use the existing badge.

Also, we're increasing the long job bonus from 50% to 75% for Lucky.

Do You Feel Lucky?

This project is a prime-finding project, so it will be available for TdP, or at least most of it. I'm not sure exactly when we'll turn it on.

The internal name is "genefer_extreme". That's what you would use in app_config.xml or app_info.xml. Should it ever be necessary, the next names would be genefer_ultimate, then genefer_ludicrous, then genefer_plaid.

Will it be added to the existing GFN badge or will there be a new badge created?

It will use the existing badge.

Also, we're increasing the long job bonus from 50% to 75% for Lucky.

Niiiiiice bonus, 👍

They say it's better to be lucky than good. But I think it's better to be good. Last year I felt lucky because I was good at aborting WUs a wingman had finished. But some people thought that was evil, not good, and put a stop to it.

So, no, I don't feel lucky this year. :(

However, if PrimeGrid comes up in either the Pentathlon or a Formula BOINC sprint this year, I want to be good again. So, which will get better points, this project with its 75% bonus or PPS Sieve?

I suppose it's PPS Sieve anyway, because I wouldn't be lucky enough to finish a WU with a wingman before whatever external challenge ended.

This project is a prime-finding project, so it will be available for TdP, or at least most of it. I'm not sure exactly when we'll turn it on.

This project is a prime-finding project, so it will be available for TdP, or at least most of it. I'm not sure exactly when we'll turn it on.

What is currently holding it from being turned on?

I'd like to offer special thanks to two people.

The first is, of course, Yves Gallot, without whom none of this would be possible. Not only did Yves write the original Genefer program, but his subsequent Genefer OCL and improved CPU programs made it possible for us to expand the Genefer searches in ways that were previously impossible.

The second person is Robish, who is simply the luckiest GFN searcher on the planet. If anyone is going to find a GFN-22, it's going to be him. Yes, this is actually the reason we said yes this time. We're going to need a ridiculous amount of luck to find a world record prime, and Robish seems to have the market cornered.

No pressure, Rob. :)

For GFN 21 - 22 and Ultimate 22, is it still required to have high DP computing power ? (HD 7950 / 7970 / Titan Black etc) or are GTX 10xx / RTX 20xx running faster ? Thank You.

From Michael's post above:

"Lucky?" tasks will be slower than normal GFN-22 tasks because neither the OCL nor OCL4 transforms can be used. All GPUs should use OCL5.

For GFN 21 - 22 and Ultimate 22, is it still required to have high DP computing power ? (HD 7950 / 7970 / Titan Black etc) or are GTX 10xx / RTX 20xx running faster ? Thank You.

Unless there's any more problems, this project is live.
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

Mike said in the chat that they're about a third longer on his 1060.

I'm currently running one on a 2080 and it should take 29-30 hours.
____________
Long live the sievers.

+ Encyclopaedia Metallum: The Metal Archives +

On a side note, Michael, do you you mind if I quote your uncle's wise words elsewhere?

Be aware that similar things have been said many times, both before and after it was said by Michael Goetz's uncle. See e.g. https://quoteinvestigator.com/2017/10/21/lottery/. /JeppeSN

GFN-22 is sieved to b=100M, so we can continue doing this search up to about 33.5 million digits. If we have to search larger numbers, we'll either need to sieve GFN-22 above 100M, or

Apparently the Genefer software can go only to b=400M, or 36 million digits. So when/if the world record prime exceeds that size, GFN-23 would be required (assuming that Genefer still has the same limitation). /JeppeSN

Be aware that similar things have been said many times, both before and after it was said by Michael Goetz's uncle. See e.g. https://quoteinvestigator.com/2017/10/21/lottery/. /JeppeSN

I never said uncle Manny invented the quote himself!

But I'm pretty sure I heard it from him in the 70's or 80's, so at least that predates many (but certainly not all) of the documented prior versions of the quote.

Trivial item: On the Preferences page, CPU should be changed to GPU since it is a GPU-only based project. ("Recent average CPU time: 0:00:00")

Eagerly waiting for some stats in relation to the average GPU elapsed time with these units.

27 "Do You Feel Lucky?" tasks are currently under work.

Now after Robish was that much hyped I looked at Robish's Badges... What the? TdP 2019 has just started yesterday and Robish ALREADY found a prime? Put him into the Lucky Project. 2 Tasks, world record prime done.

Now after Robish was that much hyped I looked at Robish's Badges... What the? TdP 2019 has just started yesterday and Robish ALREADY found a prime? Put him into the Lucky Project. 2 Tasks, world record prime done.

:) 3 luckys crunching away, fingers crossed :)

Now after Robish was that much hyped I looked at Robish's Badges... What the? TdP 2019 has just started yesterday and Robish ALREADY found a prime? Put him into the Lucky Project. 2 Tasks, world record prime done.

Mike, can you tell credits per WU with 75% bonus?
____________
92*10^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME
2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie!

Be friend with him: maybe some of his luck pass on you :)