CCminer(SP-MOD) Modded GPU kernels.

[bensam1231]

[...] WHO on the forums actually does release their miner with conditions and prices that are clearly disclosed in advance? [...]

I did, not so long ago. Works out pretty well as far as I can tell, but would likely get very messy if said miner was being sold to a large number of people. I don't know of any other examples, but I'd imagine other people have done the same.

Side note 1: Being public about the terms of the deal, and changing them after-the-fact, is utter bullshit. I'd say that's much worse than not being public about the terms but sticking to them no matter what.

Side note 2: I understand people are around here with a miner's mindset, and ROI is king. But, if you folks care for the potential of cryptocurrencies at all, maybe someday, try to pick some altcoin that you like, mine it even at a loss, grow a bag, help that community, participate, develop some service, some utility. Much more fun, and oddly enough, big profits on occasion...

I agree...  ROI is king... It's hard to open your wallet for donations when there are cobwebs at the bottom (I'm currently negative last month because of Cryptsy)... Right now a 970 makes about $1.20 mining Eth without taking out cost for power.

but then you understand, we may be reluctant to release new miners in these conditions...
I mean you can't really expect us to release for free because the conditions are bad (because it would just be bad for us as well...) hence what is expected is that we find a middle ground or a compromise which would make happy everyone. 
 
and in that case dev fee isn't either a good solution, because miners won't wait for us before dumping on a coin... (and there are a lot more miners than devs)

Not sure what you mean about a dev fee... A % based miner will be constantly mining whatever you have at your proxy setup for X algorithm. 98 shares for the miner, 2 shares for the dev.

Maybe we've always been on a different page when it comes to a miner fee? You have a miner submit 98 shares for whatever pool they're connected to and then two shares to whatever pool you want for the algorithm it's operating on. That would be a 2% fee, relatively speaking.

Or you could do it based on time... 72secs every hour.

Also not entirely certain what you mean by 'bad conditions for developers'. If you say sold your Neoscrypt miner for .1 a pop three months ago things would be quite a bit different or even right now. Kernel developers basically help set the conditions for miners. I can look at SPs miner right now and go 'hey it's not going to ROI, but I have a option of a 5% faster miner in the future for .1'. And of course SP will be improving it so it turns into a better option into the future. So something like x11 will be about 5% faster. Still not profitable, but could be.

Honestly though a % based miner is just the best solution. It rewards developers for making faster kernels constantly and gives miners big and small a chance to compete. You keep updating your miner and the amount of money you make keeps growing too. It's win-win

the problem with the dev fee, is that it collects an altcoin, and technically there are a lot more chance that people dump the coin before a dev has any chance to do it, making it a rather bad deal for the dev.

You can mine to Nicehash or a multipool and you collect BTC. That's why I mentioned it earlier. Technically you can mine to whatever the algo supports. Setting up a proxy (for you) allows you to redirect from a static address as well. You don't need to mine whatever the miner is mining.

There would be two separate stratums. Whatever the miner is mining to and then your proxy. 98 shares to the miners pool and 2 to your proxy. Or you could do it off of time.

Dev fees have been around for quite some time, there are a lot of ways to split it up if you want to. You could make a miner with two or three built in addresses as well.

Like SPs miner would get all the algos except for Neo, Neo goes to your proxy. You guys can really split it up whatever way you want after you get the fee.

[1714797421]

1714797421

[1714797421]

1714797421

[1714797421]

1714797421

[1714797421]

1714797421

[djm34]

[...] WHO on the forums actually does release their miner with conditions and prices that are clearly disclosed in advance? [...]

I did, not so long ago. Works out pretty well as far as I can tell, but would likely get very messy if said miner was being sold to a large number of people. I don't know of any other examples, but I'd imagine other people have done the same.

Side note 1: Being public about the terms of the deal, and changing them after-the-fact, is utter bullshit. I'd say that's much worse than not being public about the terms but sticking to them no matter what.

Side note 2: I understand people are around here with a miner's mindset, and ROI is king. But, if you folks care for the potential of cryptocurrencies at all, maybe someday, try to pick some altcoin that you like, mine it even at a loss, grow a bag, help that community, participate, develop some service, some utility. Much more fun, and oddly enough, big profits on occasion...

I agree...  ROI is king... It's hard to open your wallet for donations when there are cobwebs at the bottom (I'm currently negative last month because of Cryptsy)... Right now a 970 makes about $1.20 mining Eth without taking out cost for power.

but then you understand, we may be reluctant to release new miners in these conditions...
I mean you can't really expect us to release for free because the conditions are bad (because it would just be bad for us as well...) hence what is expected is that we find a middle ground or a compromise which would make happy everyone.  
  
and in that case dev fee isn't either a good solution, because miners won't wait for us before dumping on a coin... (and there are a lot more miners than devs)

Not sure what you mean about a dev fee... A % based miner will be constantly mining whatever you have at your proxy setup for X algorithm. 98 shares for the miner, 2 shares for the dev.

Maybe we've always been on a different page when it comes to a miner fee? You have a miner submit 98 shares for whatever pool they're connected to and then two shares to whatever pool you want for the algorithm it's operating on. That would be a 2% fee, relatively speaking.

Or you could do it based on time... 72secs every hour.

Also not entirely certain what you mean by 'bad conditions for developers'. If you say sold your Neoscrypt miner for .1 a pop three months ago things would be quite a bit different or even right now. Kernel developers basically help set the conditions for miners. I can look at SPs miner right now and go 'hey it's not going to ROI, but I have a option of a 5% faster miner in the future for .1'. And of course SP will be improving it so it turns into a better option into the future. So something like x11 will be about 5% faster. Still not profitable, but could be.

Honestly though a % based miner is just the best solution. It rewards developers for making faster kernels constantly and gives miners big and small a chance to compete. You keep updating your miner and the amount of money you make keeps growing too. It's win-win

the problem with the dev fee, is that it collects an altcoin, and technically there are a lot more chance that people dump the coin before a dev has any chance to do it, making it a rather bad deal for the dev.

You can mine to Nicehash or a multipool and you collect BTC. That's why I mentioned it earlier. Technically you can mine to whatever the algo supports. Setting up a proxy (for you) allows you to redirect from a static address as well. You don't need to mine whatever the miner is mining.

There would be two separate stratums. Whatever the miner is mining to and then your proxy. 98 shares to the miners pool and 2 to your proxy. Or you could do it off of time.

Dev fees have been around for quite some time, there are a lot of ways to split it up if you want to. You could make a miner with two or three built in addresses as well.

Like SPs miner would get all the algos except for Neo, Neo goes to your proxy. You guys can really split it up whatever way you want after you get the fee.

yeah well SP would get all the algo except those with fee... 

[sambiohazard]

@Slava_K +1, heck i will even run with more % at times of new releases if there is a switch for that.

i have the dev donation links with various algos to mine as a donation through the dbm project - yet not many have used it - and that goes directly through nicehash to the devs account ...

if more and more people used this kind of donation by mining - then we can all amalgamate our hasrates and make for a good donation to the devs ...

my project page - https://bitcointalk.org/index.php?topic=1089744 ...

it all needs updating and a less 'messy' op - but the info is all there ... including tnelson and stratehm donation links ...

all the links run through the stratehm stratum proxies - so not a single hash is lost through the proxy if the miner is set correctly ...

nicehash themselves have kindly whitelisted ALL the server ips so that ALL servers stay online to the nicehash system ...

give it a go and let me know if it is a system that will amalgamate / unite miners for a single cause or not ...

there is also a rotational server that donates to most of the devs on the list - but i have not made that public yet ... rotates donation time through the devs btc addresses on a 24hour basis ...

#crysx

The problem with using those as opposed to hard coded fee in the miner is that Nicehash is not profitable for me so i mine other coins mostly but if miner itself keeps mining say 2% of mins every 24 Hour like it does in Miner Control then i can mine whatever i want & still support devs. I guess for the time being i have to manually mine half an hour/day(2%) on every devs addy. that way i can support 7 devs/week. If miner had fees then i dont have to do this as it can be set up as devs want ie which algos fee go to whom etc.

[chrysophylax]

@Slava_K +1, heck i will even run with more % at times of new releases if there is a switch for that.

i have the dev donation links with various algos to mine as a donation through the dbm project - yet not many have used it - and that goes directly through nicehash to the devs account ...

if more and more people used this kind of donation by mining - then we can all amalgamate our hasrates and make for a good donation to the devs ...

my project page - https://bitcointalk.org/index.php?topic=1089744 ...

it all needs updating and a less 'messy' op - but the info is all there ... including tnelson and stratehm donation links ...

all the links run through the stratehm stratum proxies - so not a single hash is lost through the proxy if the miner is set correctly ...

nicehash themselves have kindly whitelisted ALL the server ips so that ALL servers stay online to the nicehash system ...

give it a go and let me know if it is a system that will amalgamate / unite miners for a single cause or not ...

there is also a rotational server that donates to most of the devs on the list - but i have not made that public yet ... rotates donation time through the devs btc addresses on a 24hour basis ...

#crysx

The problem with using those as opposed to hard coded fee in the miner is that Nicehash is not profitable for me so i mine other coins mostly but if miner itself keeps mining say 2% of mins every 24 Hour like it does in Miner Control then i can mine whatever i want & still support devs. I guess for the time being i have to manually mine half an hour/day(2%) on every devs addy. that way i can support 7 devs/week. If miner had fees then i dont have to do this as it can be set up as devs want ie which algos fee go to whom etc.

i agree with this mate ... but for the moment - all the proxies go through nicehash as that is the easiest way of doing it ... and the most convenient - as opposed to the most 'profitable' ...

had i a pool however ... that is a different story altogether ...

and it would NOT be a difficult thing to allow each dev to have control over each of the proxies to mine whatever the DEV wanted to mine ...

and if was a granite pool - there would be no fees or extras in that mining ... what the miner wanted to donate in algo - the dev can choose in coin - and no fees apply ...

now to find a dev that is willing to help me acheive this - first the pool - then the interface for the stratum proxy ( pfool - interested? ) for the sole control of the dev that is assigned the proxy ...

that way - 'we' miners just choose the algo we want to donate - point it at the stratum proxy - and the devs decide what they want to mine - on the pool that is designed FOR the devs ...

id be happy to support that ...

and i have thefarm to start it all off ...

#crysx

[djm34]

@Slava_K +1, heck i will even run with more % at times of new releases if there is a switch for that.

i have the dev donation links with various algos to mine as a donation through the dbm project - yet not many have used it - and that goes directly through nicehash to the devs account ...

if more and more people used this kind of donation by mining - then we can all amalgamate our hasrates and make for a good donation to the devs ...

my project page - https://bitcointalk.org/index.php?topic=1089744 ...

it all needs updating and a less 'messy' op - but the info is all there ... including tnelson and stratehm donation links ...

all the links run through the stratehm stratum proxies - so not a single hash is lost through the proxy if the miner is set correctly ...

nicehash themselves have kindly whitelisted ALL the server ips so that ALL servers stay online to the nicehash system ...

give it a go and let me know if it is a system that will amalgamate / unite miners for a single cause or not ...

there is also a rotational server that donates to most of the devs on the list - but i have not made that public yet ... rotates donation time through the devs btc addresses on a 24hour basis ...

#crysx

The problem with using those as opposed to hard coded fee in the miner is that Nicehash is not profitable for me so i mine other coins mostly but if miner itself keeps mining say 2% of mins every 24 Hour like it does in Miner Control then i can mine whatever i want & still support devs. I guess for the time being i have to manually mine half an hour/day(2%) on every devs addy. that way i can support 7 devs/week. If miner had fees then i dont have to do this as it can be set up as devs want ie which algos fee go to whom etc.

but again, we get back to the main problem with dev fee miner (ie closed source... and cannot be written under GPL license... hence need tons of rewrite to get rid of the standard everything licence under GPL).

[sambiohazard]

I dont know if this is feasible but is there a possibility of some kind of collaboration from miners, pools & other services (like nicehash did with djm) to do a crowdfund for this fee miner. This will give devs an incentive to do a rewrite. I would rather donate say 0.1 or 0.2 (1-2 month of mining for me) & 2% forever than sporadic donations that help no one.

[djm34]

I dont know if this is feasible but is there a possibility of some kind of collaboration from miners, pools & other services (like nicehash did with djm) to do a crowdfund for this fee miner. This will give devs an incentive to do a rewrite. I would rather donate say 0.1 or 0.2 (1-2 month of mining for me) & 2% forever than sporadic donations that help no one.

this goes far beyond simple crowdfunding (and seeing the little amount of participant... also why any user would pay to pay additional fee, that doesn't sound logical to me...  

Also that would represent a huge amount of work (several months full time) . I wouldn't personally have time to enter such endeavor and no matter what, I wouldn't do it until the funding had been secured.

[sambiohazard]

I dont know if this is feasible but is there a possibility of some kind of collaboration from miners, pools & other services (like nicehash did with djm) to do a crowdfund for this fee miner. This will give devs an incentive to do a rewrite. I would rather donate say 0.1 or 0.2 (1-2 month of mining for me) & 2% forever than sporadic donations that help no one.

this goes far beyond simple crowdfunding (and seeing the little amount of participant... also why any user would pay to pay additional fee, that doesn't sound logical to me...   

Also that would represent a huge amount of work (several months full time) . I wouldn't personally have time to enter such endeavor and no matter what, I wouldn't do it until the funding had been secured.

Hmm its a chiken & egg problem. I would because where thing are going right now is not good for small miners like me & i understand value of your work even if in very abstract terms. I am not mining currently as its not profitable anymore. Maybe i will do folding@home.

I hope altcoin communities come together because they need miners to be profitable & mining to be decentralized, altcoins, pools & miners.

[Grim]

What about collaborating with eg. ocminer and his suprnova pools.

AFAIK his pools are all 0% fee and he "lives" of fee donations.

So you guys release closed source miners free of charge which can only connect to his pool and ocminer introduces a fee which goes right to the miner dev account.


Fee donations still go to ocminer as before (he has to pay his servers as well).

(see, no gpl violation    )

[djm34]

What about collaborating with eg. ocminer and his suprnova pools.

AFAIK his pools are all 0% fee and he "lives" of fee donations.

So you guys release closed source miners free of charge which can only connect to his pool and ocminer introduces a fee which goes right to the miner dev account.


Fee donations still go to ocminer as before (he has to pay his servers as well).

(see, no gpl violation    )

What about collaborating with eg. ocminer and his suprnova pools.

AFAIK his pools are all 0% fee and he "lives" of fee donations.

So you guys release closed source miners free of charge which can only connect to his pool and ocminer introduces a fee which goes right to the miner dev account.


Fee donations still go to ocminer as before (he has to pay his servers as well).

(see, no gpl violation    )

gpl licence implies the code must be given to the users when a software is released (it has nothing to do with pricing or else, could be sold for a million dollars, that wouldn't violate the gpl licence; it is just a question of releasing the source)
hence if the code is given away, anyone can modify it and run it on another pool (in that current case...)

[Grim]

ah right f*** that

[djm34]

ah right f*** that

actually another alternative would be something a la opencl with precompile kernel (a priori that wouldn't violate the licensing)
but again lots of work...

[pallas]

ah right f*** that

actually another alternative would be something a la opencl with precompile kernel (a priori that wouldn't violate the licensing)
but again lots of work...

...and something which doesn't permit reverse engineering the kernel...
...and sometimes the magic is in the miner itself, not the kernel...

[djm34]

ah right f*** that

actually another alternative would be something a la opencl with precompile kernel (a priori that wouldn't violate the licensing)
but again lots of work...

...and something which doesn't permit reverse engineering the kernel...
...and sometimes the magic is in the miner itself, not the kernel...

actually what I did in my miner with fee (for vanillacoin) was to hash some stuff through keccak, then xor it into the
final hash so that if someone hack into the program, it never solves the correct hash...

[pallas]

ah right f*** that

actually another alternative would be something a la opencl with precompile kernel (a priori that wouldn't violate the licensing)
but again lots of work...

...and something which doesn't permit reverse engineering the kernel...
...and sometimes the magic is in the miner itself, not the kernel...

actually what I did in my miner with fee (for vanillacoin) was to hash some stuff through keccak, then xor it into the
final hash so that if someone hack into the program, it never solves the correct hash...

doesn't having the source of the miner be enough to figure it out?
and how did you do the fee part? if you have the sources, you can remove it.

[djm34]

ah right f*** that

actually another alternative would be something a la opencl with precompile kernel (a priori that wouldn't violate the licensing)
but again lots of work...

...and something which doesn't permit reverse engineering the kernel...
...and sometimes the magic is in the miner itself, not the kernel...

actually what I did in my miner with fee (for vanillacoin) was to hash some stuff through keccak, then xor it into the
final hash so that if someone hack into the program, it never solves the correct hash...

doesn't having the source of the miner be enough to figure it out?
and how did you do the fee part? if you have the sources, you can remove it.

well actually the source was never released and more powerful kernel were around a couple of day later, which made it totally obsolete so there wasn't anypoint to release it...  

the fee part was implemented in ccminer.cpp such that a certain %age (5% if I remember well) of all the gpu rounds were done with the dev login info and the rest with the standard miner login info. However I believe, this was done with a condition which was probably relatively easy to hack with a decompiler (pretty straight forward c++), and that was probably making the whole keccak encoding easy to bypass. (actually the %age of dev fee, dev login infos were part of the 512bit sequence (the rest was some random stuff taken here and there in the ccminer code) hashed through my initial keccak round (running on cpu only once at start), to prevent some obvious alteration of the executable...

The main hassle was that every pool had to be hardcoded as well as some other stuff...

assuming the code didn't had any alteration that key was xored to the a bunch of wirlpool constant (which had been previously xored to the original key) so that the correct solution could be found... (otherwise it was just sending wrong solutions to the pool)

[sp_]

Much easier to just build a private farm with the fastest kernals in the world. I've got 0.5 BTC for releasing a 7% faster quark kernal (sp-mod private 3). A private farm will give me more than 0.5BTC per day. Electricity here is $0.05 / KWh(Government supported industry power). and the temperature outside is below 0 C. (natural cooling).

And.. I forgot to tell you. I am not a cuda developer. I  am an assembly coder...

[djm34]

Much easier to just build a private farm with the fastest kernals in the world. I've got 0.5 BTC for releasing a 7% faster quark kernal (sp-mod private 3). A private farm will give me more than 0.5BTC per day. Electricity here is $0.05 / KWh(Government supported industry power). and the temperature outside is below 0 C. (natural cooling).

And.. I forgot to tell you. I am not a cuda developer. I  am an assembly coder...

yeah well the problem with you, is that you accept any deals...   (0.5btc from a bigfarm seems like a pretty bad deal for you...,
no matter how slow or fast is your kernel... yeah well 7% increase isn't really interesting either but hey that's me  )
thanks for stopping dumping kernel price  

[sp_]

(0.5btc from a bigfarm seems like a pretty bad deal for you...,

Unless I am owning the farm  

My private quark is still 20% faster then release 3 (sp-mod private)

My school is old school

http://www.pouet.net/

[sp_]

I have sendt version 3 of the sp-mod private to the donators/buyers. A boost in lyra2v2 and quark.

If you want to support this opensource project you should donate 0.1 BTC and recieve my private miner.

Sp-mod priv 3 on the gtx 970 (compared to release 74 (nicehashminer))

quark: +7.1%
lyra2v2: +4.9%

Sp-mod priv 3 on the gtx 750ti (compared to release 74 (nicehashminer))

lyra2v2: +5.8%
quark:  +1.8%

Only 0.1BTC guys.