1697 replies to this topic

[bauen1]

There is a link to a github page a few posts above, click it thats the KristAPI by apemanzilla

[ry00000]

No I need a lua api not a java api

[bauen1]

Thats the lua API

Edited by bauen1, 23 January 2016 - 03:03 PM.

[3d6]

I'm maintaining an updated copy of the api here https://github.com/K...dation/kristapi

[ry00000]

Yay-o! Thanks, 3d6! (You were cossacksson before, weren't you?)

[bauen1]

3d6 Is the index.php script in LegacyNote the actual source code or only parts of it?

[3d6]

ry00000, on 23 January 2016 - 05:03 PM, said:

Yay-o! Thanks, 3d6! (You were cossacksson before, weren't you?)

Yes!

bauen1, on 23 January 2016 - 05:12 PM, said:

3d6 Is the index.php script in LegacyNote the actual source code or only parts of it?

It's a little outdated, but we're doing a full rewrite right now anyways. Should be done very soon

[bauen1]

Whos joining me mining? (aka, who is k7as0j87id)

[Justy]

bauen1, on 23 January 2016 - 06:33 PM, said:

Whos joining me mining? (aka, who is k7as0j87id)

[bauen1]

justync7, on 23 January 2016 - 06:47 PM, said:

bauen1, on 23 January 2016 - 06:33 PM, said:

Whos joining me mining? (aka, who is k7as0j87id)

-snap-

Nice, what miner/hardware are you using?

Edited by bauen1, 23 January 2016 - 06:51 PM.

[Justy]

bauen1, on 23 January 2016 - 06:49 PM, said:

justync7, on 23 January 2016 - 06:47 PM, said:

bauen1, on 23 January 2016 - 06:33 PM, said:

Whos joining me mining? (aka, who is k7as0j87id)

-snap-

Nice, what miner/hardware are you using?

Ape's miner, "turbokrist" (GPU).
It isn't released yet, but it may be once the rewrite of krist is done.

[Foogles]

Argh I'd really love to get a copy of it. I get terrible numbers from CPU mining, and I have a nice GPU

[Yevano]

I've just recently had an idea of how to do a mining pool for krist. I've thought of two solutions, one of them being much more elegant than the other, however I'm not sure if the more elegant of the two is actually possible.

It goes like this: The worker computes hashes for random nonces like normal, except the hash function is different from the normal function you would use to get values lower than the work value. Let's call the normal function k, and the revised function h. Also, the work value will be w. Now, if the worker finds a nonce N such that h(N) < w, then N is sent to the pool operator. The pool operator knows about a function H such that k(H(N)) = h(N). In other words, the pool operator uses H(N) to transform the worker's useless nonce into the nonce that can be accepted by the krist server. So the pool operator would send H(N) to the krist server, receive payment, and then distribute the krist to the workers based on their accumulated shares.

Now, the problem is actually getting that pair of functions h and H so that the worker doesn't know the real nonce to send to the krist server, preventing workers from stealing money from the pool. I've thought a bit about how one would arrive at those functions, but I'm pretty stumped, and again I don't even know if it's possible. Anyways, the other solution is not too bad either as it also can't be cheated, but it requires an extra step which just makes the process not quite as clean looking. If anyone has ideas on this, let me know.

[3d6]

Yevano, on 23 January 2016 - 10:33 PM, said:

I've just recently had an idea of how to do a mining pool for krist. I've thought of two solutions, one of them being much more elegant than the other, however I'm not sure if the more elegant of the two is actually possible.

It goes like this: The worker computes hashes for random nonces like normal, except the hash function is different from the normal function you would use to get values lower than the work value. Let's call the normal function k, and the revised function h. Also, the work value will be w. Now, if the worker finds a nonce N such that h(N) < w, then N is sent to the pool operator. The pool operator knows about a function H such that k(H(N)) = h(N). In other words, the pool operator uses H(N) to transform the worker's useless nonce into the nonce that can be accepted by the krist server. So the pool operator would send H(N) to the krist server, receive payment, and then distribute the krist to the workers based on their accumulated shares.

Now, the problem is actually getting that pair of functions h and H so that the worker doesn't know the real nonce to send to the krist server, preventing workers from stealing money from the pool. I've thought a bit about how one would arrive at those functions, but I'm pretty stumped, and again I don't even know if it's possible. Anyways, the other solution is not too bad either as it also can't be cheated, but it requires an extra step which just makes the process not quite as clean looking. If anyone has ideas on this, let me know.

Require that all solutions to h are mined to a known pool address, and have the pool operator pay the miners based on how many solutions they were able to provide using the money from the solutions that actually solved blocks.

No need to trust miners not to steal funds from the pool, and no need to hide the real nonce from them. They cannot submit the same nonce with a different address.

You should PM me your Skype if you have one.

Edited by 3d6, 23 January 2016 - 10:55 PM.

[Foogles]

3d6, on 23 January 2016 - 10:52 PM, said:

-snip-
Require that all solutions to h are mined to a known pool address, and have the pool operator pay the miners based on how many solutions they were able to provide using the money from the solutions that actually solved blocks.

Wouldn't the pool operator be doing a huge brunt of the work, checking whether every random nonce submitted by a pool address is valid in this case?

[3d6]

Foogles, on 23 January 2016 - 10:54 PM, said:

3d6, on 23 January 2016 - 10:52 PM, said:

-snip-
Require that all solutions to h are mined to a known pool address, and have the pool operator pay the miners based on how many solutions they were able to provide using the money from the solutions that actually solved blocks.

Wouldn't the pool operator be doing a huge brunt of the work, checking whether every random nonce submitted by a pool address is valid in this case?

Only impressive nonces would be shown to the pool, not necessarily block solving nonces, but those that provide low hashes. It wouldn't be that much work depending on how good you want the nonces to be. You'd just have the miner send their address along with the nonce so that the pool knows to give them "shares" and when the pool finds a block, distribute the money it earns to the addresses based on how many "shares" they have accumulated and then reset everything to zero.

[Luca_S]

I am working on the same idea for a KST Mining pool, it will(Summed up) work like this(Like a concept in real BTC):

- It is based on proof of work(Shares)
- If someone mines a block the shares of every miner gets halfed
- If the pool mines a block, the shares are payed out like this: (<blockvalue>/<all shares>)*<Shares by the miner>
- A share is basically solving a block with a lower value

It is/will be written in Golang/Java(Golang -> Server, Java -> Client) and a bit of HTML and CSS, I already have a base for the TCP connection and the http server. If you want just come along and pull request to the development branch: https://github.com/Luca0208/KristPool

[Foogles]

Luca_S, on 23 January 2016 - 11:14 PM, said:

I am working on the same idea for a KST Mining pool, it will(Summed up) work like this(Like a concept in real BTC):

- It is based on proof of work(Shares)
- If someone mines a block the shares of every miner gets halfed
- If the pool mines a block, the shares are payed out like this: (<blockvalue>/<all shares>)*<Shares by the miner>
- A share is basically solving a block with a lower value

It is/will be written in Golang/Java(Golang -> Server, Java -> Client) and a bit of HTML and CSS, I already have a base for the TCP connection and the http server. If you want just come along and pull request to the development branch: https://github.com/Luca0208/KristPool

I can't wait to use it! This will give the miners that aren't as good a better shot at consistent KST.

[Luca_S]

Foogles, on 23 January 2016 - 11:17 PM, said:

I can't wait to use it! This will give the miners that aren't as good a better shot at consistent KST.

Yeah, but especially it will allow to compete against GPU miners
Edit: I have a lot of time tomorrow today(It is 00:30 here) So I can push it a lot forward, what I would need a lot by now would be someone who would work at the web interface(Important note: the code doesn't support an index.html, I will probably change that, but it might be important to know.

Edited by Luca_S, 23 January 2016 - 11:30 PM.

[bauen1]

Luca_S, on 23 January 2016 - 11:14 PM, said:

I am working on the same idea for a KST Mining pool, it will(Summed up) work like this(Like a concept in real BTC):

- It is based on proof of work(Shares)
- If someone mines a block the shares of every miner gets halfed
- If the pool mines a block, the shares are payed out like this: (<blockvalue>/<all shares>)*<Shares by the miner>
- A share is basically solving a block with a lower value

It is/will be written in Golang/Java(Golang -> Server, Java -> Client) and a bit of HTML and CSS, I already have a base for the TCP connection and the http server. If you want just come along and pull request to the development branch: https://github.com/Luca0208/KristPool

Just want to clear this up,
the https://github.com/b...KristMiningPool has been the first attempt at this.
All miners attempt to mine for the pools address and submit partially solutions (shares) to the pool and are being paid by the percentage of shares from the total shares

Edited by bauen1, 23 January 2016 - 11:30 PM.