Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 881. (Read 2347659 times)

I like this ideal of an optional donation switch in the miner.     

profits:

pentablake sp-mod: 0.01BTC per gtx 970 (24h)
quark@nichash (opensource sp-mod 72) 0,0021BTC per gtx 970 (24h)

Private pentablake miner (sp-mod 2) with overclocking:

With my optimized code, the junior gtx 970 with 1600 shaders will outperform the gtx 980ti. I can push it up to 90MHASH with heavy overclocking, but 85 is more stable. I haven't tested the latest kernal on the 980ti yet, I but would love to see it running. The pentablake sp-mod 3 is unreleased..

Your Average hashrate in the benchmarkmode is around 60MHASH, so I believe the private miner will do around 120 or more on your card. Hopefully 140MHASH with spmod release 3

You can mine on a pool here:

http://coinspool.cu.cc/workers_joincoin-pentablake

The current diff is 114

At the current diff 700 MHASH is giving around 270 joincoins a day (0.07 BTC)

The hashrate should be close to 80mh ... it needs to be implemented 
The 980ti is 4  750ti  no need for 4x the mb psu hdd  ect  I will never buy a 750ti again...there like a little cpu

With the current diff at around 100 Penta is highly profitable, but only with my fast private kernal and few miners.

run with -a penta

I didn't test it. But it would take me 2 weeks to make .3btc maybe more. On top of the almost .5btc I paid in the last year for open and one

private. If the 10,000 miners did that it would equal  5,000btc or over $1,000,000 usd. I don't have any btc to pay..ooo .12btc is all I have. And I

will have to cash that out for food and medical. There is nothing worth mining with pentablake. Wallet mining a high net hash coin = -0 profit. 

I'll benchmark mark it though....but I forget how..and it's not in the help menu.
Is it        ccminer.exe  -a pentablake  -d 0  - benchmark

I have optimized the Pentablake for ccminer  (Nvidia Maxwell gpu's)

Hashrates:

750ti: 32MHASH
970: 80MHASH ++
980ti: (I think around 140MHASH but not tested)

This is around 100% faster than the opensource miner.

7 copies have been sold so far.

The miner is for sale for 0.3 BTC

If I sell one more I will release a boost on the gtx 970 (and other compute 5.2 cards.) 85MHASH to the buyers..

No. It's broken. You can mine with the cpu, but then the ccminer hashrate of the rig will drop several MHASH. (use the -C option to improve it)


Leocoin has increased the nfactor to 16, so the speed is lower.

Does ScryptJane not work on the SP version?

I've run gobs of different settings and always get an error message like

"
[2015-10-26 04:33:33] GPU #0: cudaError 4 (unspecified launch failure) calling '
cudaMemcpyToSymbol(pdata, host_pdata, 20*sizeof(uint32_t), 0, cudaMemcpyHostToDe
vice)' (keccak.cu line 471)"

I got it working with the tpruvot version, but I'm not getting the hashrate they state I should on nicehash, and can't use their suggested settings.

Confused.

Best I could get painstakingly running threw probably 50 different settings is 850 on a 970 using -L 18 -l t13x22 , they state they are getting like 1.15KHash/s using -L 14 -l t13x28 on nicehash at https://www.nicehash.com/index.jsp?p=news&id=41

Anyway, driving me bonkers.

Thanks.

Around 4650KHASH on the low clocked cards, and 5400 with overclocking.

guys what's the exact qubit hashrate of gtx 750 ti with sp_'s modded ccminer ?

Yeah there aren't much in the way of keccak/whirl/blake coins. This is one of the unfortunate times where one of them got pumped this month (DGB) and all the GPU miners lost out on it. However, before this point they didn't make a difference. The only real scrypt coins are Litecoin and Doge.

Coinmarketcap only covers volume and not mining. It's impossible to tell based on CMC what's actually being held onto and what's being mined and dumped on a daily basis.

Yup, you'd need to be the best in the world. You're assuming that people that basically do this for a hobby that commend the pay of a full time job make the best possible kernels. Like I mentioned before, there is almost no competition in the kernel market... especially with the lack of a marketplace.


I'm sure you would... I would buy your kernel if I thought you were going to actually sell it to me... or something. We already covered this and I'm not the only one you wont sell your kernels to, same thing with DJM, and it has nothing to do with reselling them. I actually participate on the forums which should be telling in and of itself.

I'm not insulting people who don't want to do it. I don't think I directed insults at anyone at any point in time. I did say however you lost me a lot of money. And if you find that offensive, which you should, perhaps you should do something to remedy that. Like not making the rich richer and leaving all the small miners out on the curb, and there are ways of doing that which I suggested. And we can look at examples of it working too... like Claymore.

My version is still 23% faster. (tested by the cryptomining-blog.com) If you merge the rest of my kernals you get the full speed.

Quark:

My version 1.5 sp-mod 71:


your version 1.7 beta:

Scrypt and sha256 based coins are all asic now. Single algo coins like whirlpoolx, blake, keccak are all FPGA. and then we have a few algos left. The biggest problem with a 2% miner is that it has to be the fastest in the world at all times. This is very difficult. People like DJM34 and wolf0 could easily add a few percent to most of the  algos out there. Close the source and be a competetor, and then we will be out of business..

The .02% was a typo right? It would 2%. I'm not sure you can figure this out from coinmarket cap. Safer bet would be looking at the hashrate on Nicehash and take part of that. The only big ASIC alt-coin right now is Litecoin, so you can exclude that.

All the cryptos are in the shitter right now. Pretty much everything is going negative at this point. Things like Nicehash don't take into account big farms running private pools as well.

I wouldn't be worried about people ripping off the miner, there are ways of dealing with that, but they take time and money. Encryption, authentication, and verification. Those are just the brute force methods too.

A startup isn't something that makes big money. It's something that gets off the ground. Half a million per year in revenue is enough for a small company employing like 10 people with pretty decent wages. I was originally looking at something like 5-7 people as well. Guess coinmarket math is another way of looking at it.

I just had a quick look at the www.coinmarketcap.com site. Yesterday the exchanges traded bitcoins and altcoins for  $43 026 823. Bitcoin trades where 89% of the market so the altcoins had a volume of around $5 000 000 per day. According to the coinmarketcap.com site, the total market cap for bitcoins and altcoins  together is around $ 4 624 979 409. But coins like sharkcoin and Leocoin  and many of the new coins are not included, so the actual marketcap is higher.


Bitcoin marketcap: $4 150 000 000
Altcoins marketcap: $471 000 000

The inflation in the altocoin economy can be as much as 10% (the minable coins) but let's set it to 5% (Since many coins are POS only). With 5% inflation in average, the altcoins generate a mining revenue of $23 550 000 per year. $64 520 per day.

Let's say your 2% miner was the best in every altocoin algo in the world.. Then you would make $1290 per day.

but..

Many alt's can be mined by ASIC's and FPGAS. The GPU minable coins is perhaps half of that.. $600 per day..

And then somebody steal your kernal and put it on the internet for free. (perhaps with a virus)

And then you end up broke, and wasted alot of time..

Window x64 8.1. I also have a W10 x64 machine, they all perform the same.


Your entire business model depends on so called 'shitcoins'. You make a living off making kernels for them.

Putting aside the hypocrisy, there is a LOT of money that flows through GPU/CPU mineable coins. I don't think I'm underestimating anything. The trade volume is tremendous and checking out Nicehash shows that 2% would easily yield a lot of money... that's just Nicehash. The stats are all there. There aren't many businesses where potential revenue is pretty much laid out right in front of you and you can easily see how much of it you could take.

If you wanted to be a dick, you could tie it to your own pools too and make people mine on them. Which would allow you to make even more money off of it.

I think you just don't want something like this to happen because it would impact your sales (rightfully so).


Just because you don't own something doesn't mean you can't say anything. If no one uses their exchange it will die. Conversely if they're the only game in town and they rape you out the asshole (price gouge) you will also die. If there is a better method and you believe that it would help everyone out (not just you) in the long run, you definitely should say something. That's about making a proper business relationship.

We all depend on each other, willingly choking one half to build a shit palace wont help you in the long run as the fat-cats you're employed by will eventually get big enough to hire their own talent.


It's actually sorta weird that there isn't more competition in the Kernel scene and the only reason I can think this exists is because you guys keep that you sell them and how much they're sold for so hush-hush.

If Nicehash wanted to help, putting aside making their own miner, they should make a marketplace for buying and selling kernels. I think that would improve things immensely.

Yes blake512 made real perf changes (in most algos), for skein only 0.6 to 1%. But no i dont need to add all your kernels, i just added the ones which are faster on both linux and windows

Not sure if the cryptomining blog binary includes it...

Im now close to the 1.7 release, remains the SM 3.0 compatibility tests (and fix at least lyra2v2 on them).... waiting a device for real tests.