The search up to 10²¹ has started! Over 50,000 new amicable pairs have been found in the first few days.

Due to the huge size of search space, first stage of the search will only look for all amicable pairs where smaller number is of the form 3^N*...*p where N > 0 and p < 10¹¹. I expect that this will find 2-3 million new amicable pairs in a year or so.

Note that system memory requirements are higher now because all prime numbers < 10¹¹ need to be stored in memory for the search.

I'm quite excited to be a part of the team, in searching for endless numbers. Let the GPU's wind up!


--Jonny Mac
@YT Jonny Macenroe

Does this app use a different algorithm as before?

Note that system memory requirements are higher now because all prime numbers < 10¹¹ need to be stored in memory for the search.

This is useful to know but if you were able to give an example for memory needed that would be better.

For example, Boinc Manager, Disc tab shows Amicable numbers only using 1.73 Mb as I have no tasks running atm.

I allocate 20gb disc space to Boinc and have currently 15Gb free for Boinc to use.

I have 32Gb Ram and allow,
When computer is in use, use at most 50 %
When computer is not in use, use at most 90 %
Leave non-GPU tasks in memory while suspended
Page/swap file: use at most 75 %

While I guess my settings and available Ram availability is ok it would be nice to have an idea of the current potential requirements for Amicable tasks :)

---

Note that system memory requirements are higher now because all prime numbers < 10¹¹ need to be stored in memory for the search.

Resources per CPU WU are doubled - both memory and CPU-cycles. Shouldn't credit follow suit ?
Can't the allocated cores share a single copy of primes in memory ?

Attention!
There was a bug in GPU version for 10²¹ search that caused invalid results after resuming from checkpoint in some cases, so I had to fix it and restart the search from the beginning.

Anything us users need to do on our end? Abort tasks, application file, reset project?

No, old tasks will just run out soon. You might abort tasks that use v3.01 GPU application, but it's safe to just leave them running without pausing/resuming.

Ok, good to know. Also, thanks for the hard work and staying on top of things. Not many Project Admins respond as quick as you do. It's great to see!

Is there a minimum requirement for GPU RAM ?

It should work on 1GB GPUs, but I haven't tested it. You'll probably need to reduce kernel size to the minimum.

I was wondering as to why only 3192 MB were being used out of my 2 x 16 GB Radeon VIIs. Theoretically speaking, shouldn't it use all of the 16 GB it has to offer?

Current EDA for completing part two of 10' 21 is 21 Oct 2028. The second part must be a lot bigger than the first part or a lot of users have moved on. [OT] in the last month I have contributed 601,585 Credit/88 work units [/OT]

Hello! Will we have a phase 3 on the project 10^21? By the end of phase 2 will we have finished 10^21?

Naz

It's a bit too early to talk about this, but here's the search layout:

Phase 1: all odd pairs (m, n) with largest prime factor of m < 10¹¹
Phase 2: all even pairs (m, n) with largest prime factor of m < 10¹¹
Phase 3: all remaining pairs (m, n) with largest prime factor of m > 10¹¹

Phase 3 will take much longer than 1 and 2 combined.

Thank you for these details! So there is a lot of work. I will see to lead a campaign with L' Alliance Francophone in the near future in order to advance the project. Thank you for the sublime work that you provide for this fabulous project.

Regards,

Naz

Thanks for the feedback.Sergei It must just purely be drop in computing power Current speed (eqv) 670.35 x AMD Ryzen 7 3700X to the reason why the estimated completion date is currently July 18, 2026 Current runtime over the last 100 tasks is 0.81 of an hour

Hello!
I started only recently, but I've started looking into the logistics of the search. I noticed my my most recent pair is a set of even numbers with 21 digits and begin with 8s. I assume this means we're nearing the end of phase 2 (Sergei Chernykh's reply on this thread) and about to hit phase 3. Would I be right assuming that since phases 1 and 2 took about a year, phase 3 will take several years?

Your found pairs are just sorted in ascending order. The actual search order is based on factorization of the smaller number in a pair. Right now we're in 2⁵*61*... range, still several years away from the end of phase 2.

Okay, that makes a lot more sense. Thanks for the help!