Prime 6-tuplets

Take part of the competition and beat world records!
Forum rules
Want to beat a world record with Riecoin software? Let us know and share your experience! Current topics normally already cover the records you may try beating, but feel free to create new ones in the same style, especially if you found new ways to beat them. Also share your discoveries here (found a new algorithm, significant optimizations,...)!
Good Luck!
Post Reply
User avatar
Pttn
Forum Staff
Forum Staff
Posts: 222
Joined: 2021-02-18 21:32:13
Location: Switzerland
Miner Score: 54.970
Contact:

Prime 6-tuplets

Post by Pttn »

Discussion about beating the prime 6-tuplet record. Let us know about your attempts, share your optimal settings, or ask for them.

Here is a guide to configure rieMiner in order to look for records.

Current record: 2^3700 + 33888977692820810260792517451 + d, d = 0, 4, 6, 10, 12, 16 (1114 digits, 3701 bits)
By Vidar Nakling on November 8 2021.
http://www.pzktupel.de/ktuplets

Good Luck!
Riecoin developer
Stelo.xyz pool operator
User avatar
Pttn
Forum Staff
Forum Staff
Posts: 222
Joined: 2021-02-18 21:32:13
Location: Switzerland
Miner Score: 54.970
Contact:

Re: Prime 6-tuplets

Post by Pttn »

I just noticed that someone beat the 5 year old record the last month.

2^3700 + 33888977692820810260792517451 + d, d = 0, 4, 6, 10, 12, 16 (1114 digits, 3701 bits)
By Vidar Nakling on November 8 2021.

Using a software called "Sixfinder" and "based on Riecoin miners"
The record was not easy to beat, I wonder if it includes interesting speedups to deserve a new name and if the code is available somewhere if so.
If rieMiner were credited, it would have been eligible for the Bounty (of currently 4046 RIC)...
Riecoin developer
Stelo.xyz pool operator
vidarn
Posts: 1
Joined: 2021-12-08 09:22:41
Location: Norway

Re: Prime 6-tuplets

Post by vidarn »

The SixFinder program used to find the latest 6-tuplet record is based on one of the old Riecoin miners, version b14 written by dga back in 2014. I stripped out all the Riecoin and blockchain stuff and basically kept Dave Andersen's sieve as it was.

My version of the program was written in late 2014, and I have preferred to stay with that one rather than explore other variants. Over the years I have made minor changes and additions, one of them being an improved Fermat test which proved useful for numbers of the kind I am working with. (2^d + c where c is comparatively small.)

So, although I am quite impressed by the current Rieminer, it has not played any role in finding the record number.

The actual search was done on a Ryzen7 2700X and lasted two and a half calendar years, with about two years actual running time. Occasionally, I use the system for other purposes.

No, this program is presently not available in public anywhere, but I may consider a change to that.

regards,
Vidarn
Post Reply

Who is online

Users browsing this forum: No registered users and 0 guests