Topic: [ANN][DCN] Deepcoin secure hashing (CPU/GPU) New algo/ No premine/ No IPO/ PoW - page 56. (Read 182607 times)

is there a problem on supernova pool? autopay is not working and manual payment not working too since yesterday...

It's at the front page...

So this coin turns out to be another myriad. It's all about mining.

Looking good.

Small grammar change required - "mining DCN yields a much higher hash rate"

I Like new algos.Is there a gpu-miner for deephash

Yeah, now it looks better!

Thanks! Should have it finished soon. Would like to see this coin go places.

pretty awesome

Yes, indeed it is.

Side note, live preview of the redesigned website with much of the constructive criticisms taken into consideration. Still much to do. Would like to improve the header logo. Need to build out the ABOUT page. Fix all placeholder links. Add information to main sliders. Etc, etc.

Anyways, let me know your thoughts!

http://xwebhosting.org/deepcoin/

Edit: Changed the header logo quite a bit. Thoughts?

thats good for network security and also for miners.

Difficulty slowly rising.

Glad to see the increasing difficulty. A pure PoW coin with a new algo has so much more potential than other common PoS coins

difficulty goes up

{
"blocks" : 73028,
"currentblocksize" : 0,
"currentblocktx" : 0,
"difficulty" : 21.86045603,
"errors" : "",
"generate" : false,
"genproclimit" : -1,
"hashespersec" : 0,
"networkhashps" : 1256120248,
"pooledtx" : 1,
"testnet" : false
}

Ah, no, you can't make that comparison, it doesn't work that way.

That's a wee bit like saying if it takes 4 and half minutes to soft-boil an egg then soft-boiling half a dozen eggs should take just short of half an hour.

But yes, the situation is a little confused/confusing.

As far as I understand it, the absolute basics for altcoin crypto are:

256-bit crypto is appropriate for the task and good until 2040 at least
ECRYPT II's 2012 annual review of key lengths opines “Good, generic application-independent recommendation, ≈ 30 years” and that’s just for 128-bit, for 256-bit they state “Good protection against quantum computers unless Shor’s algorithm applies.”

one round of hashing is adequate to secure the blockchain
Hashing the result of a hash brings no gain in security. It seems that the otherwise-inexplicable double round of SHA256 hashing in classic Bitcoin is a simple-but-tricksy way of defending against the threat of a possible length extension attack which was was being discussed at the time but which subsequently came to naught.

The multi-hash approach was first used in SiFcoin (google translated): “Complication chain to the length of 6 different hash functions and increase bit depth to intermediate 512 - attempt to protect from further development of extremely efficient Mh / s gpu-algorithms and theory, "simple" Gh / s devices”.

Note, no mention at all made of security, just a possibility of limiting GPU advantage (and a dubious possibility at that). But somehow, this tactic to limit GPU advantage has mutated into security theatre cryptobabble: “super secure hashing” (Quark), “X15 adds 2 extra layers of hashing to the popular and already very secure X13 hashing Algorithm” (Maiacoin).

There's a pertinent stackexchange discussion Guarding against cryptanalytic breakthroughs: combining multiple hash functions. The posts by Zooko and by Thomas give a realistic appreciation of the issues.

When it's boiled down to the essence: for an altcoin, a tactic of chaining hash algos can't be proven to increase security nor can the tactic be proven not to reduce security. This latter weakness is usually considered a FAIL in crypto best practice.

The sole purpose of chaining hash functions is to protect mining from being overwhelmed by specialist hardware advantage. Any mention of enhanced security is either a misconception or simply cryptobabble and, for either reason, disheartening to see in an [ANN].

Cheers

Graham

The original question referenced some of the algos in the mix:


And indeed they are. I should have mentioned in my earlier response that they do in fact come in 256 output format as well as 512. There is a general sense that Echo means the standard 256-output format, whereas if you mean to reference the 512-output format, you should do so explicitly, i.e. Echo512. The OP's question is very valid.

However, you are asking a different question: if they're the same hashfns, does that mean x11 and Deepcoin algos are the same?

No, they are not the same. But both algos use chains of hash functions and they share some of the links, specifically the three NIST candidates Luffa, CubeHash and Echo. (Algorithm is a general term and is confusingly correct for both X11-the-chain and Luffa-the-hash-function.)

For chained-hash coins, the action usually takes place in src/hashblock.h, the header file that defines which hashfns will make up the chain and their batting order.

The algorithm basically runs: given a hash “256h” that you want to check, call Luffa512(256h), the result becomes a new “512h” hash, then call CubeHash512(512h), the result becomes the new “512h”, then call Echo512(512h) and finally the result gets chopped back to 256 and returned.

Blur the details a bit and you get

256_hash_result = trimto256(Echo512(CubeHash512(Luffa512(256_hash_to_check))))

Extend the chain by a  couple more hash functions (Fugue, Keccak, etc.) and you're in qubit and NIST5 territory, interweave them randomly but reproducibly and you're off into Quarkland; continue extending and you'll eventually reach chaincoin/x11 (same eleven hashfns, different order for two of 'em), x12, 13, 14, 15 and 17 <- the latter all pretty much just “moar hashfns”.

This time, it's hashblock.h under the lens:

Darkcoin:
https://github.com/darkcoinproject/darkcoin/raw/master/src/hashblock.h
Deepcoin:
https://github.com/Deepcoinbiz/Deepcoin/raw/master-0.8/src/hashblock.h

Despite a slight difference in coding style it's clear they are related but by no means the same.

There's an interesting-to-me sidenote at the changepoint “62c36”: Deepcoin's chain function is appropriately labelled “Hash5” whereas Darkcoin's chain function is labelled “Hash9” (not 11), apparently a fossilised remnant of Mr. Duffield's design process. Other coin devs appear to have honoured this, at least X15Coin has https://github.com/X15COIN/X15Coin/blob/master/src/hashblock.h#L78

Personally, I'm very cautious of statements about the improbability of ASICs for NIST candidates, whichever round ...



(From http://rijndael.ece.vt.edu/sha3/ - note, the free giveaway of SHA-3 ASICs has finished.)

I recommend a visit to Graz Technical University's SHA-3 Zoo where you check up on things like http://ehash.iaik.tugraz.at/wiki/SHA-3_Hardware_Implementations


Cheers

Graham

+1

Yes It will take time to develop an ASIC miner for a newer algo, during this time we can enjoy  

Now the ASIC manufacturers are hating you for this line

maybe i'm wrong but in theory you can develop asic device with any algorithm, it's very expensive, so I think all new coins protected from such devices anyway

anti-ASIC ability: just change the algorithm every year