CCminer(SP-MOD) Modded GPU kernels.

[tsiv]

Hi All
Being watching this thread with interest the past could of weeks & decided to move over from ageing old AMD's to Nvidia and build my first 750ti rig. which is now up and running  
I've got to ask tsiv how are you getting 5700 from the 750ti? mine are only doing 4600  
wDragon

Wrote a faster version of the lyra kernel. Magic.

[sp_]

15% more power and 25% more hash. That's a good 10% increase. I will buy that miner for 0.1 BTC if you will sell it. You can probobly sell 20-30 copies and earn a few btc's

[pallas]

15% more power and 25% more hash. That's a good 10% increase. I will buy that miner for 0.1 BTC if you will sell it. You can probobly sell 20-30 copies and earn a few btc's

Or, you could make a pledge address and when you reach the amount you like, opensource it.
I think that djm34 also has a faster lyra2rev2 and sp_'s got a faster quark: it would be interesting to keep a thread on those closed source optimization to understanbd where it's all going.
I might have some in the future as well, but I'd really like that it all ends in opensource at the end of the day. Let's not make nvidia a new amd ;-)

[wdragon]

15% more power and 25% more hash. That's a good 10% increase. I will buy that miner for 0.1 BTC if you will sell it. You can probobly sell 20-30 copies and earn a few btc's

Wow! Yeah I'll donate 0.1 BTC for a copy too. 

[fenomenhaa]

15% more power and 25% more hash. That's a good 10% increase. I will buy that miner for 0.1 BTC if you will sell it. You can probobly sell 20-30 copies and earn a few btc's

Wow! Yeah I'll donate 0.1 BTC for a copy too. 

i'm ready to donations too,it's quiete good improvement.

[pokeytex]

15% more power and 25% more hash. That's a good 10% increase. I will buy that miner for 0.1 BTC if you will sell it. You can probobly sell 20-30 copies and earn a few btc's

Wow! Yeah I'll donate 0.1 BTC for a copy too. 

I would also donate .1 for it.  I would donate for a faster Spreadx11 miner as well.
-pokeytex

[coinut]

count me in for .1 btc also 

[djm34]

15% more power and 25% more hash. That's a good 10% increase. I will buy that miner for 0.1 BTC if you will sell it. You can probobly sell 20-30 copies and earn a few btc's

Or, you could make a pledge address and when you reach the amount you like, opensource it.
I think that djm34 also has a faster lyra2rev2 and sp_'s got a faster quark: it would be interesting to keep a thread on those closed source optimization to understanbd where it's all going.
I might have some in the future as well, but I'd really like that it all ends in opensource at the end of the day. Let's not make nvidia a new amd ;-)

sound interesting my private should be slower (based on 750ti hashrate), must be shuffling too much (was written at the time of the old lyra, but with the new lyra, it probably makes more sense to do the other way... need to have a look)

[scryptr]

DONATIONS--

Count me in for improved Lyra2v2!  There are some dedicated donators that follow this thread.  Currently, my 750ti cards are getting 4.5+ Mh/s with SP_ release dot 62.  VertCoin was over .0004 BTC yesterday.....

--scryptr

[bensam1231]

Yeah same... This is why it'd be great if there was just a percentage you paid into the miner. Everyone pays and reaps the rewards. If you guys are publicly disclosing your numbers, I'm sure there are big miners who have their own coders on their payroll.

Looking at it another way, if you release the miner with a 2% fee, how long would it take for miners who use it to reach your 'goal' for a donation bounty? Even if someone takes it and guts the fee, and assuming all miners are bad people and they use the fee less miner regardless of knowing it's supporting your work, you probably could make the donation amount before it happens.


May I suggest as a test case, releasing a version of CCminer with a 2% fee and 10% greater hashrate for Lyra2v2 and Quark. Then go from there? You aren't giving out all your cards. You can see how fast the miner is gutted and see how many people continue using the miner after a 'fee free' version pops up. It would give you some idea of the income you could generate with a fee and whether or not it's worth your time to pursue such an endeavor.

[djm34]

Yeah same... This is why it'd be great if there was just a percentage you paid into the miner. Everyone pays and reaps the rewards. If you guys are publicly disclosing your numbers, I'm sure there are big miners who have their own coders on their payroll.

Looking at it another way, if you release the miner with a 2% fee, how long would it take for miners who use it to reach your 'goal' for a donation bounty? Even if someone takes it and guts the fee, and assuming all miners are bad people and they use the fee less miner regardless of knowing it's supporting your work, you probably could make the donation amount before it happens.


May I suggest as a test case, releasing a version of CCminer with a 2% fee and 10% greater hashrate for Lyra2v2 and Quark. Then go from there? You aren't giving out all your cards. You can see how fast the miner is gutted and see how many people continue using the miner after a 'fee free' version pops up. It would give you some idea of the income you could generate with a fee and whether or not it's worth your time to pursue such an endeavor.

it was already done with vanillacoin, but it is a mess to code considering you have to send the fee somewhere...
ie a pool, then you have to add several pools (then restrict usage to those pool) etc...

probably better to do that with an existing (non optimized version)   rather than releasing new code to "test"... (nice try...  )

Also why 2% ? This would have to be scaled over the %age optimization and not some random number.
say something gives 100% increase (ie double the hashrate), I am pretty sure that it is worth 10% fee at least for a while

nb: the easiest way to impose a fee, would probably to find some arrangement with pool admin and make binaries working only on that pool with part of the fee going to the dev (like it was done for some prime number algos)

[zTheWolfz]

Not much but I guess every little bit help. Wow it charges a fee every time you send now?
I don't recall it doing that before on BTC, just every now and then if I recall right..

- @sp_ : Transaction ID: 6f3ef10c412df4da57adf4d9d2c685d4c0dfdee610a2440bd2ff9acbc410f1c5-000
- @djm34 : Transaction ID: 438f780019f9788797a58b982b46575daa47e5b6f5b3db8cfe9e254ed723ba57-000
- @pallas : Transaction ID: 88b404d6b9036c6b123d192f54d8accb52f1da347a5a5f5855dc3c229d648b7c-000

Thanks for your support. I have added more hash for the gtx 960 (quark.) the other cards are also abit faster.
If you can build yourself, please try release 62+ (github)

Keccac512 with less instructions

If someone could point me to a good simple link for compiling for Lunix or windows I'd be glad to give it a shot.
I spent about 3 hours last night hunting and trying different walk threw, but most were not up to date ones.
So not much luck so far. I'm using Linux Mint 17.2 64bit for this rig, but only has a 660Ti in it at the moment.
Most of my rigs are open air so not much trouble changing video cards around for testing.

[joblo]

If someone could point me to a good simple link for compiling for Lunix or windows I'd be glad to give it a shot.
I spent about 3 hours last night hunting and trying different walk threw, but most were not up to date ones.
So not much luck so far. I'm using Linux Mint 17.2 64bit for this rig, but only has a 660Ti in it at the moment.
Most of my rigs are open air so not much trouble changing video cards around for testing.

Linux compilation is pretty straightforward as described in the readme file assuming you have
all the Linux development tools installed and nvidia cuda tools 6.5. Windows requires Visual Studio 2013.
If you're environment is setup correctly both work out of the box.

If you explain what you tried and where you got stuck I can give more specific tips.

[pallas]

nb: the easiest way to impose a fee, would probably to find some arrangement with pool admin and make binaries working only on that pool with part of the fee going to the dev (like it was done for some prime number algos)

This has already been done multiple times by srcxxx (now dashminer).
I instead am supporting the idea of a multipool (based on yaamp) which sends a small percent to the most active developers.
People would simply mine on it like they usually do.

[sp_]

Submitted a 20khash increase in lyra2v2 on the 750ti.  +(0.4%)

I got alot left to reach the speed of the tsiv miner...

[sp_]

-Faster quark
-Faster lyra2v2
-Some Bugfixes and improvements done by Trent Nelson. (Faster and more stable miner exit)


1.5.63(sp-MOD) is available here: (31-08-2015)

https://github.com/sp-hash/ccminer/releases/

The sourcecode is available here:

https://github.com/sp-hash/ccminer

[sp_]

If someone could point me to a good simple link for compiling for Lunix or windows I'd be glad to give it a shot.
I spent about 3 hours last night hunting and trying different walk threw, but most were not up to date ones.
So not much luck so far. I'm using Linux Mint 17.2 64bit for this rig, but only has a 660Ti in it at the moment.
Most of my rigs are open air so not much trouble changing video cards around for testing.

I have buildt release 63 for you to test.

[majinboo]

If someone could point me to a good simple link for compiling for Lunix or windows I'd be glad to give it a shot.
I spent about 3 hours last night hunting and trying different walk threw, but most were not up to date ones.
So not much luck so far. I'm using Linux Mint 17.2 64bit for this rig, but only has a 660Ti in it at the moment.
Most of my rigs are open air so not much trouble changing video cards around for testing.

I have buildt release 63 for you to test.

Looks like colors are broken on r63 but it seems to add around 100 khs on lyra2v2 for gtx 750.

[zTheWolfz]

If someone could point me to a good simple link for compiling for Lunix or windows I'd be glad to give it a shot.
I spent about 3 hours last night hunting and trying different walk threw, but most were not up to date ones.
So not much luck so far. I'm using Linux Mint 17.2 64bit for this rig, but only has a 660Ti in it at the moment.
Most of my rigs are open air so not much trouble changing video cards around for testing.

I have buildt release 63 for you to test.

Nice work!   Right at 300kh/s above what I was getting for Quark.
Edit: X11 5240kh/s, about the same as before.
Testing lyra2rev2 now 5857kh/s unsure of my numbers before on this, neoscrypt up next.

[bensam1231]

Yeah same... This is why it'd be great if there was just a percentage you paid into the miner. Everyone pays and reaps the rewards. If you guys are publicly disclosing your numbers, I'm sure there are big miners who have their own coders on their payroll.

Looking at it another way, if you release the miner with a 2% fee, how long would it take for miners who use it to reach your 'goal' for a donation bounty? Even if someone takes it and guts the fee, and assuming all miners are bad people and they use the fee less miner regardless of knowing it's supporting your work, you probably could make the donation amount before it happens.


May I suggest as a test case, releasing a version of CCminer with a 2% fee and 10% greater hashrate for Lyra2v2 and Quark. Then go from there? You aren't giving out all your cards. You can see how fast the miner is gutted and see how many people continue using the miner after a 'fee free' version pops up. It would give you some idea of the income you could generate with a fee and whether or not it's worth your time to pursue such an endeavor.

it was already done with vanillacoin, but it is a mess to code considering you have to send the fee somewhere...
ie a pool, then you have to add several pools (then restrict usage to those pool) etc...

probably better to do that with an existing (non optimized version)   rather than releasing new code to "test"... (nice try...  )

Also why 2% ? This would have to be scaled over the %age optimization and not some random number.
say something gives 100% increase (ie double the hashrate), I am pretty sure that it is worth 10% fee at least for a while

nb: the easiest way to impose a fee, would probably to find some arrangement with pool admin and make binaries working only on that pool with part of the fee going to the dev (like it was done for some prime number algos)

What sort of incentive is there to use a fee based miner when the current version is free, besides it's doing the right thing? And are you really bitching about having to do a little bit of work to earn money?It really isn't that hard to setup a pool and you know it, you just don't want to have to do the tiniest bit extra work to make it happen (which shows the whole reason we're stuck where we are right now).

Because people see a 10% fee and they hate you. Then they tear out your fee to spite you. It's supposed to be a agreeable amount for everyone. 2% is low enough where people will accept it and use it, even if a 'fee free' alternative is released. 2% is affordable. If a 'fee free' version comes out people will see 10% additional profit, instead of 2%. It then becomes about being competitive with other miners are earning more money and expanding much faster instead of doing the right thing.

Also a percent is a percent. If you have a 100% faster kernel, you will double how much you're making (if nothing else changes) through the donation. It scales. It isn't a flat fee. Continued improvements will reward the developer with increased profits. Ideally improvements such as the difference between quark public and quark private wont be that big, the only reason they are right now is because people have been holding back.

nb: the easiest way to impose a fee, would probably to find some arrangement with pool admin and make binaries working only on that pool with part of the fee going to the dev (like it was done for some prime number algos)

This has already been done multiple times by srcxxx (now dashminer).
I instead am supporting the idea of a multipool (based on yaamp) which sends a small percent to the most active developers.
People would simply mine on it like they usually do.

You can use whatever sort of pool you want to make it happen. A donation pool is really no different then a donation system right now. I guess some people prefer it, but if it came down to it, I'd just send some BTC. It's more inconvenient then just sending BTC if someone is more inclined. A fee is there because everyone then does it and there is straight income from the product. The developer improves their kernels, they see a improvement in the amount of income they get. It's incentive to continue to support the product and the end user doesn't need to do anything.

I'm pretty certain if a donation pool was setup, the current situation will not change and we'll go back to the same system where developers hold back because they feel they aren't getting adequate support for their miner.