sp_ (OP)
Legendary
 Offline
Activity: 2898
Merit: 1087
Team Black developer
|
 |
August 26, 2015, 09:56:12 PM
Last edit: August 26, 2015, 10:14:56 PM by sp_ |
|
Submitted an optimalization in bmw-256.
Do bmw with less assembly instructions
lyra2v2 +10KHASH on the 750ti.
|
|
|
|
|
|
|
|
|
The network tries to produce one block per 10 minutes. It does this by automatically adjusting how difficult it is to produce blocks.
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
|
|
skunk
|
|
August 26, 2015, 11:37:41 PM |
|
Submitted an optimalization in bmw-256.
Do bmw with less assembly instructions
lyra2v2 +10KHASH on the 750ti.
lyra2/lyra2REv2.o: In function `scanhash_lyra2v2':
tmpxft_00002d40_00000000-3_lyra2REv2.cudafe1.cpp:(.text+0x6f8): undefined reference to `bmw256_cpu_hash_32(int, unsigned int, unsigned int, unsigned long*, unsigned int*, unsigned long)'
collect2: error: ld returned 1 exit status
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2814
Merit: 1091
--- ChainWorks Industries ---
|
|
August 27, 2015, 12:51:09 AM |
|
so any ideas sp? ... what would you prefer? ... just donations? ...
#crysx
I can include a 1% fee, and keep it opensource. I just need to refork tvpruvot's work, and add my modded kernals. The 1.6.6 version supports mining to multiple pools. Then I just need to optimize every algo a few percent, and everybody would use my optimized kernals... If it keeps you happy and interested. 2% is still fine dude. 1% SP, 1% split between other contributors... I think people might even do 3%, but higher then that and people will start getting rowdy. Make sure you note it at the beginning and when it's mining shares for you though so people can easily see what's happening. If you really want to push this you can add a bit of your meaty private kernel to it to beef it up and make people want to swap (if they're hashing more then 2% faster it's essentially a wash). My private quark kernal is 16% faster than the opensource, and I am currently expanding my farm. then possibly this model is the best model to use in this space ... open source miner is at least 10% less the private miner - which you can sell privately ... full stop ... whether i had one machine mining or one hundred machines mining - i would pay for the private miner everytime becuase of two things ... 1 - it helps with the incentive to you sp with the development of the miner ... 2 - it helps with the mining potential of the machines i use to mine ... i have paid for miners in the past that i cannot use - all because its a good cause to donate ... you have released windows only miners that is impossible for me to use - yet i still pay for the miner as a donation source ... apart from the donation mining that we have contributed - that makes for a few donations ... if everyone paid for a private miner - you would develop it further ... the other devs would be donated to also ... and lets include ALL the other here that donate to you directly ... add it all up ... still not enough? ... then we need to do something better ... this goes for ALL the devs that have donated THEIR time and effort into making this ( and many other miners ) optimized and better ... so possibly this is the best model for the time being until we can think of something better ... if the private miner is up for 'sale' or we need to donate for access to it - let us know ... #crysx |
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2898
Merit: 1087
Team Black developer
|
|
August 27, 2015, 03:35:30 AM |
|
lyra2/lyra2REv2.o: In function `scanhash_lyra2v2':
tmpxft_00002d40_00000000-3_lyra2REv2.cudafe1.cpp:(.text+0x6f8): undefined reference to `bmw256_cpu_hash_32(int, unsigned int, unsigned int, unsigned long*, unsigned int*, unsigned long)'
collect2: error: ld returned 1 exit status
fixed |
|
|
|
|
sambiohazard
|
|
August 27, 2015, 03:44:20 AM |
|
Why not go for a three tiered plan?
CCMiner Free - not fully optimized, totally free.
CCMiner Lite - 10% faster than Free, mines at 1-2% fee to sp_. (I will take this one)
CCMiner Pro - 20% faster than Free for x BTC, comes with some features that large farms want & need.
Adjust the % of optimization to what you think is fair for your input & what you are getting. I guess many will use "Lite" version just as its 10% faster than free
|
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2814
Merit: 1091
--- ChainWorks Industries ---
|
|
August 27, 2015, 04:50:35 AM |
|
Why not go for a three tiered plan?
CCMiner Free - not fully optimized, totally free.
CCMiner Lite - 10% faster than Free, mines at 1-2% fee to sp_. (I will take this one)
CCMiner Pro - 20% faster than Free for x BTC, comes with some features that large farms want & need.
Adjust the % of optimization to what you think is fair for your input & what you are getting. I guess many will use "Lite" version just as its 10% faster than free
and what of the other devs that commit optimizations / inclusions / rewrites of the various algos and implementations of 'pulls' that others do for ccminer-spmod? ... dont forget - this fork is sp's fork - but it is also a conglomeration of work from so many other authors ... it looks good on paper ( or txt here in this case ) but i honestly believe its impractical ... i have a farm - would i pay a certain amount for the 'pro' version? ... i can ... but why not use the smaller versions in that case? ... or use the versions that tpruvot puts out? ... or djm34's? ... or klaust's? ... or ... see what i mean? ... besides - that looks like a microsoft model of doing things ... no tanx from my end ... #crysx |
|
|
|
|
sambiohazard
|
|
August 27, 2015, 06:36:17 AM |
|
Why not go for a three tiered plan?
CCMiner Free - not fully optimized, totally free.
CCMiner Lite - 10% faster than Free, mines at 1-2% fee to sp_. (I will take this one)
CCMiner Pro - 20% faster than Free for x BTC, comes with some features that large farms want & need.
Adjust the % of optimization to what you think is fair for your input & what you are getting. I guess many will use "Lite" version just as its 10% faster than free
and what of the other devs that commit optimizations / inclusions / rewrites of the various algos and implementations of 'pulls' that others do for ccminer-spmod? ... dont forget - this fork is sp's fork - but it is also a conglomeration of work from so many other authors ... it looks good on paper ( or txt here in this case ) but i honestly believe its impractical ... i have a farm - would i pay a certain amount for the 'pro' version? ... i can ... but why not use the smaller versions in that case? ... or use the versions that tpruvot puts out? ... or djm34's? ... or klaust's? ... or ... see what i mean? ... besides - that looks like a microsoft model of doing things ... no tanx from my end ... #crysx hmm but its really not easy to donate unless its a coin like VTC, small miners have to wait a lot before donating BTC. If miner mines to devs its easy, may be devs can come to some agreement among themselves on how to split things up. As for MS model, i guess its good to have features you want & pay for it. This is different from MS in the sense that miners can have a say in what feature they want nothing forced on them, and there is always free version so no one is forced to pay either. Right now we have this very same arrangement with sp's private kernels available which are 20% faster. We can come up with all kind of plans but it really depends on how much profit miners are making & how fairly they want to share in the economy. Its more about ethics than anything else. We have seen that what happens when profits are not shared (AMD miner model). We are also seeing same in real economy & stock markets that when profits are not shred/reinvested its bad for everyone. So if people are not dumb/greedy & short sighted & want to cannibalize their own profits, then they will pay or market will self adjust by going towards either development stagnation or private miners. |
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2898
Merit: 1087
Team Black developer
|
|
August 27, 2015, 06:41:37 AM |
|
The people who donate 0.1BTC to my account (drk/vtc is also accepted) gets a version of the Spreadcoinminer with full sourcecode. The miner is more than 10% faster than the opensource miner, so it's a real advantage when the difficulty is low.
|
|
|
|
chrysophylax
Legendary
Offline
Activity: 2814
Merit: 1091
--- ChainWorks Industries ---
|
|
August 27, 2015, 09:29:25 AM |
|
Why not go for a three tiered plan?
CCMiner Free - not fully optimized, totally free.
CCMiner Lite - 10% faster than Free, mines at 1-2% fee to sp_. (I will take this one)
CCMiner Pro - 20% faster than Free for x BTC, comes with some features that large farms want & need.
Adjust the % of optimization to what you think is fair for your input & what you are getting. I guess many will use "Lite" version just as its 10% faster than free
and what of the other devs that commit optimizations / inclusions / rewrites of the various algos and implementations of 'pulls' that others do for ccminer-spmod? ... dont forget - this fork is sp's fork - but it is also a conglomeration of work from so many other authors ... it looks good on paper ( or txt here in this case ) but i honestly believe its impractical ... i have a farm - would i pay a certain amount for the 'pro' version? ... i can ... but why not use the smaller versions in that case? ... or use the versions that tpruvot puts out? ... or djm34's? ... or klaust's? ... or ... see what i mean? ... besides - that looks like a microsoft model of doing things ... no tanx from my end ... #crysx hmm but its really not easy to donate unless its a coin like VTC, small miners have to wait a lot before donating BTC. If miner mines to devs its easy, may be devs can come to some agreement among themselves on how to split things up. As for MS model, i guess its good to have features you want & pay for it. This is different from MS in the sense that miners can have a say in what feature they want nothing forced on them, and there is always free version so no one is forced to pay either. Right now we have this very same arrangement with sp's private kernels available which are 20% faster. We can come up with all kind of plans but it really depends on how much profit miners are making & how fairly they want to share in the economy. Its more about ethics than anything else. We have seen that what happens when profits are not shared (AMD miner model). We are also seeing same in real economy & stock markets that when profits are not shred/reinvested its bad for everyone. So if people are not dumb/greedy & short sighted & want to cannibalize their own profits, then they will pay or market will self adjust by going towards either development stagnation or private miners. agreed ... i would definitely pay for features in the miner for thefarm that i would like to be used in thefarm ... optimizations are another thing that - if demonstrated - can pull more mining power to the devs ... paying for features isnt really an ms model as such - i was just likening the model to them - as they are a greedy bunch ... but you are spot on when it comes to paying for features - as has been stated earlier ... i have mentioned much earlier for a lean towards a 'new' miner that has the functionality and form of ccminer with the 'stats-candy' of sgminer ... this of course means a rewrite - or a bare bones start - of a miner that can do this ... sgcc if you will - for arguments sake ... all plugable and modularized so that optimizations dont have to be written into the miner - but plugged in ... of course - even then - we are still in a very similar situation with 'what will the devs get out of it' scenario ... id like to believe in a donation community and people would be willing to give a little to get a lot ... but instead - some give and a lot dont - to get a lot ... none of us can afford a development stagnation in the crypto market - none of us ... as sp suggested - private miners of the likes os spreadminer v9 that sp has is as little as 0.1btc WITH the source ... which did me just fine as i could compile and USE the miner ... so as for smaller miners waiting a while before they donate btc - then the links that i have forthcoming very shortly would help a great deal - as they can point their miners to the algo they want for the dev that they want to donate to - and mine FOR the dev ... all through nicehash currently ... so as small as their hashrate maybe - they can add up to a lot for the devs if everyone pitches in a little hashrate - especially at the same time  ... anyway - back to work - otherwise i wont get this sorted in the next 7 hours ... #crysx |
|
|
|
bensam1231
Legendary
Offline
Activity: 1750
Merit: 1024
|
|
August 27, 2015, 09:35:39 AM |
|
Why not go for a three tiered plan?
CCMiner Free - not fully optimized, totally free.
CCMiner Lite - 10% faster than Free, mines at 1-2% fee to sp_. (I will take this one)
CCMiner Pro - 20% faster than Free for x BTC, comes with some features that large farms want & need.
Adjust the % of optimization to what you think is fair for your input & what you are getting. I guess many will use "Lite" version just as its 10% faster than free
and what of the other devs that commit optimizations / inclusions / rewrites of the various algos and implementations of 'pulls' that others do for ccminer-spmod? ... dont forget - this fork is sp's fork - but it is also a conglomeration of work from so many other authors ... it looks good on paper ( or txt here in this case ) but i honestly believe its impractical ... i have a farm - would i pay a certain amount for the 'pro' version? ... i can ... but why not use the smaller versions in that case? ... or use the versions that tpruvot puts out? ... or djm34's? ... or klaust's? ... or ... see what i mean? ... besides - that looks like a microsoft model of doing things ... no tanx from my end ... #crysx hmm but its really not easy to donate unless its a coin like VTC, small miners have to wait a lot before donating BTC. If miner mines to devs its easy, may be devs can come to some agreement among themselves on how to split things up. As for MS model, i guess its good to have features you want & pay for it. This is different from MS in the sense that miners can have a say in what feature they want nothing forced on them, and there is always free version so no one is forced to pay either. Right now we have this very same arrangement with sp's private kernels available which are 20% faster. We can come up with all kind of plans but it really depends on how much profit miners are making & how fairly they want to share in the economy. Its more about ethics than anything else. We have seen that what happens when profits are not shared (AMD miner model). We are also seeing same in real economy & stock markets that when profits are not shred/reinvested its bad for everyone. So if people are not dumb/greedy & short sighted & want to cannibalize their own profits, then they will pay or market will self adjust by going towards either development stagnation or private miners. There are a lot of different ways you could present a product and probably no one model is right. I know % based mining fees are the easiest for all involved. If you do a 'paid' version you have to setup a actual way of accepting payment (unless they're just going to do manual transactions with paper receipts). You can split % based mining among contributors as well. Also agree about the AMD mining conundrum. Good example of all the profit going to a handful of people. I definitely think developers should be fairly compensated, but all the burden (or profit) shouldn't be put on one group of individuals. I've pushed for something like this a few different times because I don't want to see development stop and I think it's quite fair (as long as the percent isn't jacked up a lot). As I mentioned earlier, if CCMiner had a API, they could easily adopt MultiMiner to be a frontend and add stats... It's opensource. |
I buy private Nvidia miners. Send information and/or inquiries to my PM box.
|
|
|
chrysophylax
Legendary
Offline
Activity: 2814
Merit: 1091
--- ChainWorks Industries ---
|
|
August 27, 2015, 10:04:36 AM |
|
Why not go for a three tiered plan?
CCMiner Free - not fully optimized, totally free.
CCMiner Lite - 10% faster than Free, mines at 1-2% fee to sp_. (I will take this one)
CCMiner Pro - 20% faster than Free for x BTC, comes with some features that large farms want & need.
Adjust the % of optimization to what you think is fair for your input & what you are getting. I guess many will use "Lite" version just as its 10% faster than free
and what of the other devs that commit optimizations / inclusions / rewrites of the various algos and implementations of 'pulls' that others do for ccminer-spmod? ... dont forget - this fork is sp's fork - but it is also a conglomeration of work from so many other authors ... it looks good on paper ( or txt here in this case ) but i honestly believe its impractical ... i have a farm - would i pay a certain amount for the 'pro' version? ... i can ... but why not use the smaller versions in that case? ... or use the versions that tpruvot puts out? ... or djm34's? ... or klaust's? ... or ... see what i mean? ... besides - that looks like a microsoft model of doing things ... no tanx from my end ... #crysx hmm but its really not easy to donate unless its a coin like VTC, small miners have to wait a lot before donating BTC. If miner mines to devs its easy, may be devs can come to some agreement among themselves on how to split things up. As for MS model, i guess its good to have features you want & pay for it. This is different from MS in the sense that miners can have a say in what feature they want nothing forced on them, and there is always free version so no one is forced to pay either. Right now we have this very same arrangement with sp's private kernels available which are 20% faster. We can come up with all kind of plans but it really depends on how much profit miners are making & how fairly they want to share in the economy. Its more about ethics than anything else. We have seen that what happens when profits are not shared (AMD miner model). We are also seeing same in real economy & stock markets that when profits are not shred/reinvested its bad for everyone. So if people are not dumb/greedy & short sighted & want to cannibalize their own profits, then they will pay or market will self adjust by going towards either development stagnation or private miners. There are a lot of different ways you could present a product and probably no one model is right. I know % based mining fees are the easiest for all involved. If you do a 'paid' version you have to setup a actual way of accepting payment (unless they're just going to do manual transactions with paper receipts). You can split % based mining among contributors as well. Also agree about the AMD mining conundrum. Good example of all the profit going to a handful of people. I definitely think developers should be fairly compensated, but all the burden (or profit) shouldn't be put on one group of individuals. I've pushed for something like this a few different times because I don't want to see development stop and I think it's quite fair (as long as the percent isn't jacked up a lot). As I mentioned earlier, if CCMiner had a API, they could easily adopt MultiMiner to be a frontend and add stats... It's opensource. it seems that the percentage for mining is the main way that a dev can get his / her portion of the development donation ... but as sp iterated earlier - it wouldnt take much for someone to refork and remove such code ... the amd model is a fairly simple one - but as you have stated - in the hands of an elite few ... im not against that - except that monopolies are formed that way and that is quite detrimental to the mining community ... and im not against it because the few devs that DO develop the optimizations and miners do release their code and optimizations when they feel that it is ready for the community to take ... in any case - it really isnt this model of opensource availability - so its not really a working model that we can really look at ... i think that a community of users that GIVE as they receive is a better way of donating to the developers than forcing percentages and keeping miners private ... BUT - unfortunately - not everyone GIVES in order to use these miners and optimizations ... if that was the case - we would not even be discussing this ... is ALL miners gave a little once a week - then the devs would have a LOT of incentive to continue development ... which is why donating hashrate - no matter how small - may be all the help the devs need ... and moreso - all the donation the smaller miners CAN give for the time being ... #crysx |
|
|
|
lawrencelyl
Member
Offline
Activity: 94
Merit: 10
|
|
August 27, 2015, 11:54:46 AM |
|
It's beer time... Some VTC beer to sp_: f1239503cb973c08865ba8759bda0de6e5228fa70f561da81352aa3a82b7013f  |
|
|
|
|
|
sambiohazard
|
|
August 27, 2015, 12:48:45 PM |
|
@crysx
I appreciate your work with the donation links. I was wondering why are you not creating a multipool. This way you can split the fees between yourself & devs or may be give any excess donations to devs. All the miners who post here & your own farm can mine on your pool instead of nicehash so we have enough hashrate to be profitable. This way everyone wins. yaamp is now opensource so you can use or modify that code. Also if this turns out to be profitable then someone from community can customize the pool code to add features. Also i heard that sp_ is creating a pool or something like that, may be you 2 can work together. What say everyone?
|
|
|
|
|
|
|
|
kama
|
|
August 27, 2015, 05:52:18 PM |
|
great news ! i was waiting a service like this |
|
|
|
|
zTheWolfz
|
|
August 27, 2015, 05:58:47 PM |
|
What does this -p d=0.002 do?  |
|
|
|
|
|
skunk
|
|
August 27, 2015, 06:15:03 PM |
|
What does this -p d=0.002 do? it's the desired share difficulty you can set as password parameter on some multipools or on nicehash |
|
|
|
|
zTheWolfz
|
|
August 27, 2015, 06:33:02 PM |
|
What does this -p d=0.002 do? it's the desired share difficulty you can set as password parameter on some multipools or on nicehash So a difficulty of 0.002 is better on some algo/hardware than 0.001 it uses at default when nothing is set? |
|
|
|
|
|
skunk
|
|
August 27, 2015, 06:36:43 PM |
|
What does this -p d=0.002 do? it's the desired share difficulty you can set as password parameter on some multipools or on nicehash So a difficulty of 0.002 is better on some algo/hardware than 0.001 it uses at default when nothing is set? no. read https://www.nicehash.com/index.jsp?p=faq#faqs10 |
|
|
|
|
zTheWolfz
|
|
August 27, 2015, 06:47:24 PM |
|
What does this -p d=0.002 do? it's the desired share difficulty you can set as password parameter on some multipools or on nicehash So a difficulty of 0.002 is better on some algo/hardware than 0.001 it uses at default when nothing is set? no. read https://www.nicehash.com/index.jsp?p=faq#faqs10Got it now thanks. |
|
|
|
|
|
