Post reply

Posted by: JoeMagi

« on: 10. September 2014., 20:07:26 »

Insert Quote

XMG - Coin of The Magi

Freenode IRC: #magi

(Hybrid PoW/PoS-II)

______________________________________________________

Proof of Wok

Hash: M7M - only minable via CPU
M7M manipulates certain mathematical functions to frustrate GPU mining;
Block time: 3 minutes;
Difficulty dependent block rewards;
1st phase PoW: primary mining with maximum 500 XMG/block;
2nd phase PoW: maximum 10 XMG/block, 7% annual reduction.

______________________________________________________

Proof of Stake-II

Introduce a new staking mechanism: PoS-II, designed resistant to "rich-getting-richer";
Minimum stake coin age : 2 hours;
Dynamic coin age: maximum 30 days, subject to change;
Dynamic annual interest: depends on network weight, maximum 5%;
Block time: 1.5 minutes;
Block hash: SHA256
Maturity: 100
PoS-II initiated after two weeks of PoW

(Total coins PoW+PoS: 25 million)

# Magi - the most energy efficient and secured system

Magi is designed to be energy efficient on the whole, "superficially" through many dynamic adjustment mechanisms. Present block generation is through hybrid PoW/PoS. The general idea is to reduce PoW mining through GPU (or probably ASIC), with the assistance of secured PoS-II.
The PoW mining is done through a new forged M7M algo, mining via CPU only (resistant to GPU & ASIC);
A mechanism is yet to be introduced to suppress PoW while increase PoS when big miners come into the network;
We plan to add a further mining mechanism, giving additional rewards to CPU miners; further details yet to be deployed.

# M7M - a CPU mining algo
M7M is based on prior work done by XCN, the M7 algorithm. XCN has done well on choosing cryptographically secure hashing functions; M7 hashes the multiplication of seven hashes by using SHA-256. The M7 algorithm has been proven a success. There was a period when M7 is only minable through CPU. Unfortunately, as of this time both Nvidia and AMD miners have been implemented, that makes M7 like any other algorithms.

Under this scheme, we have carried out modifications to the M7 algo. The basic idea is that GPU does bad work on manipulating complicated mathematical functions compared to CPU. The resultant algorithm significantly increases the difficulty of implementing GPU miners. This idea and our implementation could be very promising for future use in CPU coins. The nature characteristic resists to GPU and hence ASIC too. The additional advantage is that the algorithm is not memory hard which means a lot of energy consumption during mining processes. The energy wasting in Bitcoin mining has now been a world-wide concern.

# PoS-II
One of the major innovations appears to be the PoS-II feature. Yet, it is still under the test, we dedicate that it would be "rich-getting-richer" resistant. We will release detail information later on.

# Magi is smart
We provide smart adjustment throughout the coin design, as mentioned in #1; full description will be released later.

Some initial plans

Coin launch & swap, ~~we will hold back the release of source code for a few days~~;
Initiate PoS, more info will be released;
Wallet enables the Tor anonymity;
Deploying a mining mechanism additional to regular PoW/PoS mining, won't be completely new but rather based on what had in the past (with improved functioning probably, goal is to reduce power usage on the whole);
Deploying anonymous features;
More plans coming while we will make sure PoW/PoS runs properly at this moment.

Tor (anonymity network) directs Internet traffic through a worldwide network which allows to conceal a user's location. Although digitalcoins like Bitcoin can be used "anonymously", any transactions are still traceable through the blockchain. Using the Tor network makes it even more difficult for one to track any other peers on the network. We schedule to deliver the Tor wallet once PoS is initiated (about two weeks).

http://www.hashharder.com/m7/magi

# Coin swap
Magi is not deigned to be a completely new coin; rather, it is a continue of our prior Magicoin project: https://bitcointalk.org/index.php?topic=548368.0, based on Scrypt PoW. Being seeing the advent of tough time for Scrypt coin as well as lack of innovations at the launch, we started to seek creative features to be merging with this coin. We've been around since April, 2014, and simply don't want to give up any efforts we've tried; among paramount factors are the community we've maintained, for example, several facebook groups are managed by our team and people who liked Magicoin. Since we started from there, we are intent to continue from there, just this simple. Thanks to this, all of services including social network, webpage, faucets and the community will continue which will save a lot of time for us to create from the scratch.

Totally 4.5 million Magicoins were minted, as such the same amount of XMG will be used for coin swap purpose. The old Magicoin network will cease to mint new coins after block 75,000. This allows us to maintain the Magicoin network "forever"; no deadline is set for the coin swap. Ask us anytime to proceed as you wish.

We realize that there could be "a lot" coins left without swapping for whatever reasons. Over time, we will release slowly these unswapped coins to the public, completely FREE (no IPO/ICO etc), for example, giveaway to Bitcointalk users with Jr. Member and above.

I am sure you will question the stake we could gain by holding so many coins. We guarantee this would not happen (PoS disabled), because we speak TRUST on us (LOL); ok, seriously, having so many coins sitting in a wallet won't generate any interests because of the PoS-II mechanism. Details coming soon once the coin is launched.

# Phase 1 - primary PoW: 5.5 million XMG to be minted
We decide to release 5.5 million coins into public during Phase 1 through PoW mining, in addition to 4.5 million for coin swap purpose. This is to attract new miners into this new designed coin. All of the coins would not be mined in 2 days, 7 days or any short time period, but rather depending on the network difficulty. Block reward will be ~~null~~ 10 XMG if the network difficulty is low enough; it increases with higher difficulty, with a maximum of 500 XMG when diff = 150, then decreases exponentially with increasing difficulty. We estimate it may take a month (with full maximum reward) to years to reach the end of Phase II.

PoS will be initiated after two weeks PoW mining. According to the interest rate, the amount of coins generated will be much less than PoW during Phase 1 if PoW mining hard.

# Phase 2 similar amount of PoW and PoS coins minted
The maximum PoW block reward is 10 XMG, and the reward decreases with the difficulty (square root). This nearly generates similar amount to the PoS to be minted annually. However, the ratio of PoW over PoS is still subject to change, once we implement a mechanism to adjust PoW mining reward (worse case is to disable it) when the network sees significant hash impacts. Our intention, one of our further plan, is to gradually suppress PoW mining through GPU etc and hence to maintain the mining process more energy efficient in total and secured network.

XMG is based on a new M7M algo. M7M is based on modifications to the M7 algorithm. The following investigation qualitatively shows the energy efficiency of M7. Similarly, M7M will have advantage in saving energy as well.

Shown below is a data table, basically some hash rates according to testnet tests of mining though an Inteal Core i5-4670 @ 3.40GHz:

Algo 1 core 2 cores 3 cores 4 cores
Fugue: 43 - 58 C (390KH/s      50 - 63 C (780KH/s)      54 - 68 C (1750KH/s)      65 - 71 C (1900KH/s)
Whirlpool: 44 - 60 C (144-156KH/s)      50 - 66 C (300KH/s)      54 - 69 C (420K)         61 - 74 C (590KH/s)
Groestl:      43 - 60 C (140KH/s)      50 - 68 C (270KH/s)      56 - 73 C (410K)         65 - 75 C (540KH/s)
FRESH:      42 - 60 C (89KH/s)         49 - 67 C (175KH/s)      55 - 72 C (255KH/s)      63 - 74 C (340KH/s)
NIST5:      44 - 60 C (60KH/s)         51 - 67 C (155KH/s)      56 - 72 C (230KH/s)      65 - 77 C (305KH/s)
M7:      44 - 59 C (68KH/s)         51 - 66 C (133KH/s)      56 - 71 C (200KH/s)      63 - 76 C (260KH/s)
X11:         44 - 60 C (36KH/s)         51 - 66 C (70KH/s)         56 - 72 C (110KH/s)      65 - 76 C (160KH/s)
X13:         44 - 61 C (25KH/s)         51 - 66 C (49KH/s)         55 - 71 C (73KH/s)         61 - 75 C (95KH/s)
HVC:         42 - 59 C (19KH/s)         49 - 64 C (39KH/s)         53 - 69 C (55KH/s)         62 - 72 C (88KH/s)

This ranking is roughly similar to the investigation of power usage done by cryptomining-blog: http://cryptomining-blog.com/3036-power-usage-of-geforce-gtx-750-ti-with-various-crypto-algorithms/. The Fugue is most energy efficient; Whirlpool is similarly efficient based on users' experience, while Groestl appreas to be an exception. We should notice that only one fugue256 hash function is used in Fugue, in comparison with chained (X11 etc) / parallel (M7, HVC) configuration of multiple functions. Intuitively, it is quite reasonable that a miner will stay cool if the work to be done is not that computationally hard. From this point, using one hash function should be computationally easy compared to seek target hash through multiple functions. Although without any apparent proofs, we could speak that chained algo, in which hash is generated and passed to the next one by one (i.e., one input generates the output), behaves the same way of one-hash-function algo from the security point of view. For this reason, the algo with multiple inputs should be more secured, which is exactly what the M7 and HVC do.

Dutch (by 111magic): https://bitcointalk.org/index.php?topic=740276.0
Indonesian (by feryjhie): https://bitcointalk.org/index.php?topic=760497.0
Chinese (by baby222): https://bitcointalk.org/index.php?topic=763187.0
Portuguese (by libivan): https://bitcointalk.org/index.php?topic=763774
Spanish (by zarton): https://bitcointalk.org/index.php?topic=767083.0
Hebrew (by hustler100): https://bitcointalk.org/index.php?topic=768497.0
Swedish (by blacksails): https://bitcointalk.org/index.php?topic=771526.0
Turkish (by Professor James Moriarty): https://bitcointalk.org/index.php?topic=771373.0
Italian (by Yza_azY): https://bitcointalk.org/index.php?topic=771928.0
French (by torpedo): https://bitcointalk.org/index.php?topic=771859.0
Indian (by alanswami): https://bitcointalk.org/index.php?topic=772324
Russian (by fsb4000): https://bitcointalk.org/index.php?topic=774235
German (by blubberli) https://bitcointalk.org/index.php?topic=776461.0

PM for possible opportunities;

News:

SCF Portal Stats

Members

Stats

Users Online

Post reply

Topic Summary

Posted by: JoeMagi

Posted by: JoeMagi

Posted by: devnullius

Posted by: JoeMagi

Enter your email address to receive daily email with 'SCforum.info - Samker's Computer Forum' newest content: