[ANN][GAP] Gapcoin - Prime Gap Search - New Math Algo - CPU / GPU - Zero Premine

[YomKi]

I'm looking for some simple code (e.g. Python, Go, C, Perl, Java, etc.) to scrape blocks to JSON / text.  I don't want the transactions, I want the block data.  E.g. the "Raw Block" data from https://chainz.cryptoid.info/gap/block.dws?688423.htm.

gapcoin.org went down but was still doing the scraping for a while, but that stopped as well after a while and nobody has submitted any records for Gapcoin for almost a year.  If I have a good way to scrape the block data I can do regular submissions.

Gapcoin is down to 748 records (from 1102 at the beginning of the year and 1614 at its peak).  The best Gapcoin merit is currently the 29th highest known, but it's possible there is a higher merit sitting in the blockchain.

[YomKi]

Interesting item posted to Hacker News, a paper published at October's Fields Medal Symposium in Toronto: A Schr¨odinger Equation for Solving the Riemann Hypothesis (PDF).

The author has previously published this on vixra, which doesn't bode well.  He isn't on the Fields Symposium program nor is he an invited speaker.  He was an attendee.

Some interesting discussion of the Bender-Brody-Müller paper the Moxley one assumes:

• https://math.stackexchange.com/questions/2211278/riemann-hypothesis-is-bender-brody-m%C3%BCller-hamiltonian-a-new-line-of-attack
• https://arxiv.org/pdf/1704.02644.pdf  ("Conclusion: As attractive this idea looks, it does not hold when checking the analysis part of the problem."

There is some fundamental disagreement there, with Bellissard indicating the BBM paper is heavily flawed while Moxley runs with it.

[gjhiggins]

Thank you for that informed analysis, especially the stackexchange link, I shall be following that up.

doesn't bode well

I half suspected as much. I'm old enough to remember Arnold Arnold (I can only find this one reference https://www.amazon.co.uk/Geometric-Solution-Fermats-Last-Theorem/dp/B001OR8FJQ) from the PR ruckus of the time.

Cheers

Graham

[Webcelerator]

Found in the purse a few gaps - is it traded where?

[emrgohff]

Why is it that the difficulty (many Chinese to mine) is increasing a lot and in return to low?

What is up?

[emrgohff]

/Satoshi:0.9.5/   ?

Where is this version available?

[emrgohff]

GapCoin Blockchain Explorer
up to block 692945


Network Clients seen in the last 24 hours
Version, Protocol & Network Share

/Satoshi:0.9.5/   

node list   

88.98.87.243


Where is this version available?

[gjhiggins]

/Satoshi:0.9.5/   ?
Where is this version available?

Sorry, didn't see this msg til just now ... it's preliminary work, linux source only atm: https://github.com/gjhiggins/gapcoin

Cheers

Graham

[emrgohff]

OK

Is this what it is doing to increase the difficulty with no return on price.

Am I not understanding the difficulty / price?

Difficulty increases, equal or lower price

[gjhiggins]

Is this what it is doing to increase the difficulty with no return on price.

The instance is not mining, it's sole activity is to respond to getblock API calls made by a bit of Python intended as a response to:

I'm looking for some simple code (e.g. Python, Go, C, Perl, Java, etc.) to scrape blocks to JSON / text.

https://gist.github.com/gjhiggins/358c08af544e10e7ee03ff8b9b124005

Cheers

Graham

[emrgohff]

Ok, I got it

[PeterTheGrape]

I'm looking for some simple code (e.g. Python, Go, C, Perl, Java, etc.) to scrape blocks to JSON / text.  I don't want the transactions, I want the block data.  E.g. the "Raw Block" data from https://chainz.cryptoid.info/gap/block.dws?688423.htm.

gapcoin.org went down but was still doing the scraping for a while, but that stopped as well after a while and nobody has submitted any records for Gapcoin for almost a year.  If I have a good way to scrape the block data I can do regular submissions.

Gapcoin is down to 748 records (from 1102 at the beginning of the year and 1614 at its peak).  The best Gapcoin merit is currently the 29th highest known, but it's possible there is a higher merit sitting in the blockchain.

Just to be clear, as long as the blockchain exists somebody with the skill can extract all the gaps? And also was wondering that about the other math coins that are similarly neglected primecoin, riecoin and there are others whose names I forget. The problem with submitting them late is only that the work might get done twice? 

[YomKi]

Thanks Graham.  My main issue is getting a URL that returns the JSON text for the blocks.  Parsing after that is easy with various methods, but I'm not sure what the right URL is.

Just to be clear, as long as the blockchain exists somebody with the skill can extract all the gaps?

To the best of my knowledge, yes.

For instance, I can manually use a link like https://chainz.cryptoid.info/gap/block.dws?694002.htm, select the raw block, and see that at 25 Dec 2017 20:09:52 GMT, someone generated the gap

4794 2886118820041672361179922121599608532750340047263831260627518277405293606690385 02244433

that has a merit of 24.08050173.  Most of the blocks are not records.  The record gap for 4794 has a merit of 29.39 for instance, so this block isn't interesting.

And also was wondering that about the other math coins that are similarly neglected primecoin, riecoin and there are others whose names I forget.

No idea.

The problem with submitting them late is only that the work might get done twice?

Submitting them late just means nobody knows about them.  Since the date is encoded in the block, there ought not be any issue with someone reporting them as if they found them.  Gapcoin chooses the P1 randomly in a large enough space that it is *extremely* unlikely anyone would select the same P1.

There are a few different methods being used for current gap finding efforts:

• Exhaustive search.  This is looking for true record gaps, which means it started at 2 and went up from there.  It's intensely computationally expensive.  Tomás Oliveira e Silva ran a distributed project from 2005 to 2012 that got to 4e18 using years of work on hundreds of cores.  Interestingly, the computational result was used in Helfgott's 2013 proof of the Odd Goldbach Conjecture.  Recently the PGS team at mersenneforum have used a different method to extend this and after about 9 months have brought this to 10e18.  The number of records per computational effort is very small, however these are true Minimal Gaps -- once found the record is permanent, as no earlier gaps of that size exist.

• Gapcoin.  For relatively small P1 values (84-347 digits), choose a random small range, sieve out small multiples, then run Fermat tests to find gaps.  While each step is efficient and fast, it's rather inefficient at finding record gaps.  It's basically rapidly throwing darts while blindfolded and being spun around -- the only way to get more darts in the target is to throw faster.

• Primorial methods.  Gaps are far more common at multiples of primorials without some small divisors, e.g. numbers of the form N * p# / k with k a small square free number.  So if one looks at increasing values of N * 191#/30, for instance, using efficient methods for finding the previous and next primes around that point, one can find record gaps many times times faster than the gapcoin method.  That is the method used by most other searchers and is what holds all but 3 of the highest merits (those three being from the exhaustive search).  There are some minor variations -- Hans Rosenthal in 2017 did searches with a fixed large N and instead varied k.  Using the dart analogy from before, this is throwing darts while aiming at the target.  The darts are thrown a lot slower, but since they're all thrown in the direction of the target rather than randomly around the room, more of them result in high results.

There may be other methods.  I don't know how Helmut Spielauer finds results (he holds most of the records from 1400-3000), and while searchers like Michiel Jansen use the N*p#/k form, it's not known what software or algorithm is used.

[PeterTheGrape]

...
There are a few different methods being used for current gap finding efforts:

• Exhaustive search.  This is looking for true record gaps, which means it started at 2 and went up from there.  It's intensely computationally expensive.  Tomás Oliveira e Silva ran a distributed project from 2005 to 2012 that got to 4e18 using years of work on hundreds of cores.  Interestingly, the computational result was used in Helfgott's 2013 proof of the Odd Goldbach Conjecture.  Recently the PGS team at mersenneforum have used a different method to extend this and after about 9 months have brought this to 10e18.  The number of records per computational effort is very small, however these are true Minimal Gaps -- once found the record is permanent, as no earlier gaps of that size exist.

• Gapcoin.  For relatively small P1 values (84-347 digits), choose a random small range, sieve out small multiples, then run Fermat tests to find gaps.  While each step is efficient and fast, it's rather inefficient at finding record gaps.  It's basically rapidly throwing darts while blindfolded and being spun around -- the only way to get more darts in the target is to throw faster.

• Primorial methods.  Gaps are far more common at multiples of primorials without some small divisors, e.g. numbers of the form N * p# / k with k a small square free number.  So if one looks at increasing values of N * 191#/30, for instance, using efficient methods for finding the previous and next primes around that point, one can find record gaps many times times faster than the gapcoin method.  That is the method used by most other searchers and is what holds all but 3 of the highest merits (those three being from the exhaustive search).  There are some minor variations -- Hans Rosenthal in 2017 did searches with a fixed large N and instead varied k.  Using the dart analogy from before, this is throwing darts while aiming at the target.  The darts are thrown a lot slower, but since they're all thrown in the direction of the target rather than randomly around the room, more of them result in high results.

There may be other methods.  I don't know how Helmut Spielauer finds results (he holds most of the records from 1400-3000), and while searchers like Michiel Jansen use the N*p#/k form, it's not known what software or algorithm is used.

This was raised earlier, but doesn't Dana Jacobson pretty well get most of the records? FX had tried to contact them, maybe without success. Would it be practical to raid that person's github and see if we could find a dev to incorporate what that person uses? 

[gjhiggins]

Thanks Graham.  My main issue is getting a URL that returns the JSON text for the blocks.  Parsing after that is easy with various methods, but I'm not sure what the right URL is.

The node API is on 127.0.0.1:<RPCPORT> where <RPCPORT> defaults to 31397.

I dumped the postgresql database: https://minkiz.co/library/gapblocks.sql.gz

I crudely interpret the following as “nothing to report of late” ...

gapblocks=# select distinct(merit) as merits, gaplen, height from blocks where gaplen > 20000 order by merit desc limit 20;
   merits    | gaplen | height |
-------------+--------+--------+
 34.65926775 |  27666 | 483909 |
 33.66212713 |  26858 | 581982 |
 31.46506560 |  25108 | 503937 |
 31.46147878 |  25120 | 379773 |
 31.25067310 |  24950 | 520208 |
 31.22273371 |  24910 | 479509 |
 31.02379813 |  24758 | 543334 |
 30.89132028 |  24646 | 402139 |
 30.54500687 |  24380 | 510941 |
 30.50778491 |  24348 | 519574 |
 30.46456046 |  24318 | 516329 |
 30.44425897 |  24300 | 413719 |
 30.44194759 |  24300 | 361844 |
 30.41753388 |  24282 | 413770 |
 30.29834124 |  24178 | 448132 |
 30.22634743 |  24130 | 602737 |
 30.08368413 |  24016 | 516371 |
 29.91801267 |  23878 | 379253 |
 29.84590646 |  23828 | 515423 |
 29.75992059 |  23756 | 555220 |
(20 rows)

Cheers

Graham

[YomKi]

The node API is on 127.0.0.1:<RPCPORT> where <RPCPORT> defaults to 31397.

Ah, thanks.  I don't have gapcoin running but I understand now, and can do this in the future.

I dumped the postgresql database: https://minkiz.co/library/gapblocks.sql.gz

Perfect.  I got this, parsed, and there are 540 new records (albeit most are breaking own records).  Including the new number 1 and number 5 highest merits!  I'll send them in right away.

[PrimeObsession]

Dear fellow Gapcoin enthusiasts, it turnes out that Gapcoin project, despite its design limitations, is still capable of setting the highest merit record!

gapcoin-cli verifymessage GJtsALGXvATz8DEakokLDQRa9dF49AFt7F Hwgf2IaiuFScMw9754ygul3WtAwst7y4+SJK47L64XoNaVzthMmc+6uIKLPQqsNNgN105wzfSVoIFf0p0PrW3T4= "Andrey Balyakin dedicates this prime gap of merit 41.938784, which he found on October 2nd, 2017, to Gapcoin's 3rd anniversary."

gapcoin-cli verifymessage GQv89XPznFS3fFi4J3LwQSRs5sy21EoPQJ IOwRTQ4ne1cbz9i37HpEAOTcWJKQqK+eC7hwELEV0DMt8gNf5FyDUBpawivzASuzVETARr7F3Zg11AUw5ewmQ0E= "Andrey Balyakin dedicates this prime gap of merit 36.929239, which he found on October 19th, 2017, to Gapcoin's 3rd anniversary."

[gjhiggins]

Just to be clear, as long as the blockchain exists somebody with the skill can extract all the gaps?

Pretty much, yes. There's less technical effort involved in repeatedly polling the API and dumping the results in a database than it is to write a coin-specific block/chain parser.

And also was wondering that about the other math coins that are similarly neglected primecoin, riecoin and there are others whose names I forget. The problem with submitting them late is only that the work might get done twice?  

According to DOACC (which I ceased updating in Sept 2015), there are four algos using some aspect of prime number calculation to implement a proof-of-work hash but only three are pertinent to this context:

* Prime 6 - Bytecent - closed source, uninteresting. Self-description asserts that the PoW,  whilst using primes, makes no contribution to the maths at all.
* Cunningham primes: Pnut, CDNCoin, Bernankecoin, Primecoin, Datacoin
* Prime gap finding as Proof of Work: Gapcoin
* Prime constellations: Riecoin

Only Primecoin, Datacoin, Riecoin and Gapcoin have active networks, the others are in stasis.

But basically, yes, something can be done about establishing a means of collating and reporting any reportable results...

In a not-entirely-unrelated strand of investigative work, I've been taking advantage of the blockchain being structured as an acyclic directed graph and mapping the JSON output (incl. transactions) to RDF, stored, presented and made available for querying via SPARQL, courtesy of Fuseki. ATM, it's animated via a Pyramid (Python web app framework) prototype web app (which I have whimsically and irritatingly named ACME “A Cryptocurrency Metadata Explorer”) which, in essence, offers a rather tractable and readily-extensible block explorer.

I have a couple of pre-production instances implemented for off-the-beaten-track alts, Slimcoin (http://tessier.bel-epa.com:5064/) and Datacoin (http://tessier.bel-epa.com:5059/). I hope you'll forgive screenshots - I've  yet to upgrade the hardware to support the additional demand on compute resources, so the response time varies from “slow”, through “glacial” to “never” (so, if you're actually trying the links, please be patient ...

Landing page for Slimcoin showing blockchain stats, summary details of the last few blocks and a chart showing the current ratio of Slimcoin tripartite minting:


Datacoin version:



Datacoin's API also offers:

listprimerecords <primechain length> [primechain type]
and
listtopprimes <primechain length> [primechain type]

From which I can create a presentation (atm, crudely dumped into a single page until I figure out some meaningful UI approach).



Gapcoin's API offers:

listbestprimes amount (min merit)
and
listprimerecords merit

which will also permit of a more accessible presentation.


Progress on Gapcoin ACME ...



The landing page polls the RPC API for the info (on each page reload, it's pre-production).

I've made some progress on mapping the Gapcoin blockchain to RDF ...

gjh@chrome /opt/acme/gapcoin-acme/acme $ ./scripts/blocknotify-catchup.sh
/opt/acme/gapcoin-acme/acme/scripts /opt/acme/gapcoin-acme/acme
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
110000
120000
130000
140000
150000
160000
170000
180000
190000
200000
210000
220000
230000
240000
250000
260000
270000
280000
290000
300000
310000
320000
330000
340000
350000
360000
370000
380000
390000
400000
410000
420000
430000
440000
450000
460000
470000
480000
490000
500000
510000
520000
530000
540000
550000
560000
570000
580000
590000
600000
610000
620000
630000
640000
650000
660000
670000
680000
690000
E
======================================================================
ERROR: test_catchup (__main__.Test_BlockNotifyCatchUp)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "blocknotify-catchup.py", line 33, in test_catchup
    fp.write(self.g.serialize(format="nt").decode('utf-8'))
  File "/opt/acme/gapcoin-acme/lib/python3.5/site-packages/rdflib-4.2.2-py3.5.egg/rdflib/graph.py", line 943, in serialize
    serializer.serialize(stream, base=base, encoding=encoding, **args)
  File "/opt/acme/gapcoin-acme/lib/python3.5/site-packages/rdflib-4.2.2-py3.5.egg/rdflib/plugins/serializers/nt.py", line 32, in serialize
    stream.write(_nt_row(triple).encode(self.encoding, "_rdflib_nt_escape"))
MemoryError

----------------------------------------------------------------------
Ran 1 test in 103949.692s

FAILED (errors=1)

My mistake, of course 32Gb of RAM is insufficient, the mapping results in 20 to 30 triples per block, * 690,000 = “many” (and 103949.692s / 3600 = 28.874914444444446 hours)

I'll just set it to map 100k blocks at a time.

Oh, ACME also shows txs ...



and presents a (cyclopean) view of the network ...



Seasons' Greetings. More details after the w/end.

Cheers

Graham

[M0ndialu]

Gap coin is on fire !!!   1 $ is coming (low supply=big price)

https://www.cryptopia.co.nz/Exchange/?market=GAP_BTC

[bonosurgut]

Gap coin is on fire !!!   1 $ is coming (low supply=big price)

https://www.cryptopia.co.nz/Exchange/?market=GAP_BTC

any hot news or just pump ?