CCminer(SP-MOD) Modded GPU kernels.

[antantti]

We need more solominers or a working stratum implementation.. . 3 hours without a single solo block.

Epsylon? Can you do it, and add pascalcoin to yiimp.

240 000 pascal coins are mined every day. ($180 000). If nanopool has 90% of the nethash, they earn $3240 everyday with the 2% fee.


Nanopool (AMD only) has most of the hashrate

sp_, is that fuck_nano_pool you? 

[Schleicher]

I guess it is you Schleicher :-D
Seeing the progress in the repo..maybe share with us which bug you have^^ :-D

We need to beat of the shit from AMD fanboys. :-D

Well, it's the pascal_cpu_hash function.
It's either launch failure, or illegal memory access. Or it just exits, without error message.
Probably the same bug is in my bitcoin algo.

[sp_]

sp_, is that fuck_nano_pool you?  

Hehe. No. My rigs start with GG. Just found a block. Around 10 blocks yesterday, but today lower..

[chrysophylax]

I guess it is you Schleicher :-D
Seeing the progress in the repo..maybe share with us which bug you have^^ :-D

We need to beat of the shit from AMD fanboys. :-D

Well, it's the pascal_cpu_hash function.
It's either launch failure, or illegal memory access. Or it just exits, without error message.
Probably the same bug is in my bitcoin algo.

i bet you are compiling with cuda75 ...

you will see the fix ...

btw - gcc6x doesnt compile any of the miners due to the syntax of the miner code - it needs changing for the newer gcc6 and above ( fedora is already at gcc63 ) ...

anyone up for the job? ... skype me if so ...

#crysx

[bensam1231]

without the possibility to connect to a pool it's useless, solo is dead, well maybe not if you have 100 rig like in your case...

I got 10 blocks yesterday in 24 hours with 50 rigs(1060 and 1070 cards). So you should be able to hit a block a day with 5 modern rigs.

Spent about 8 hours with roughly 75GH two days ago after Nano went up and didn't hit a single block. As always when it comes to hash, the more you have, the more blocks you'll hit and less variance you'll have.

If you have five rigs you'll more then likely just not hit a block at all due to the low odds and becomes insanely unprofitable very fast.

The standard miner doesn't feed all the gpus with the same work. Just look at the found nonces: they are not the same.
BTW my miner is faster, I will publish it as soon as possible.

Yes, that's because the normal miner changes the name XXXXXX01, 02, 03. HOWEVER if you have more then one rig, you either need a different private key to mine to or a different miner name. I'm sure if you're just opening another instance it's duplicating work. Someone should look at the nonces for the same GPU.


Pallas said he fixed CPU utilization, SP if you haven't, that's still a big issue.

[bensam1231]

In other news we got 2-3 retards with large farms that are bouncing around the small alt coins over the last couple weeks, completely destroying the difficulty, sitting on it for like six hours and doing it all over again to a different coin. Sad thing is there isn't just one of them so you can't flee to a different coin.

Just another example of eth babies that have no idea how mining works and are using a profit auto-miner. The difficulty adjusts after a couple blocks are found and any gained profits from switching to those coins are lost as difficulty adjusts and revenue ends up less then Zcash, in addition to ruining things for other people.

You don't make more money doing this. You end up making less. You're a dumbass, learn to calculate your own profits... at the very least adding them up. Just because something is more profitable doesn't mean it will be that way when your fat ass ends up on the coin. Awesomeminer doesn't figure everything out for you.

Stick to Ethereum and Zcash.

[sp_]

Yes, that's because the normal miner changes the name XXXXXX01, 02, 03. HOWEVER if you have more then one rig, you either need a different private key to mine to or a different miner name. I'm sure if you're just opening another instance it's duplicating work. Someone should look at the nonces for the same GPU.

I have fixed it in the sp-mod #1 take a look at the .bat file in the archieve.

The d parameter will perform a d modulo #of gpus.

So if you have 6 gpus's in the rig.
d0-d5 will run the first thread.
d6-d11 will run the second thread.
d12-d17 will run the third thread.
d18-d23 will run the forth thread.

That's why you need to specify 24 gpus in the proxy if you want to run a rig with 6 gpu's with 4 threads.

Pallas said he fixed CPU utilization, SP if you haven't, that's still a big issue.

My version has a 6% PCIE bus load. I think the midstate should be moved to const mem. Pallas will release his version soon, then we will know what he did.

[7enuke]

SP_, your miner on 2 Thread give me
1 x 1060 3g / gpu +200 mem -500
1 x 750Ti 1g / gpu + 100 mem -500
summary 900 Mhash

2 x GTX750 2G / gpu + 100 mem -500
1 x GTX750Ti 2G / gpu + 100 mem -500
Summary 700 Mhash

[m1n1ngP4d4w4n]

I see alot of bubbling around pascal in this thread, but in my understanding, it's reserved for now for people with gigantic farms since there are no pool version of the miner right ?

[bensam1231]

Yes, that's because the normal miner changes the name XXXXXX01, 02, 03. HOWEVER if you have more then one rig, you either need a different private key to mine to or a different miner name. I'm sure if you're just opening another instance it's duplicating work. Someone should look at the nonces for the same GPU.

I have fixed it in the sp-mod #1 take a look at the .bat file in the archieve.

The d parameter will perform a d modulo #of gpus.

So if you have 6 gpus's in the rig.
d0-d5 will run the first thread.
d6-d11 will run the second thread.
d12-d17 will run the third thread.
d18-d23 will run the forth thread.

That's why you need to specify 24 gpus in the proxy if you want to run a rig with 6 gpu's with 4 threads.

Pallas said he fixed CPU utilization, SP if you haven't, that's still a big issue.

My version has a 6% PCIE bus load. I think the midstate should be moved to const mem. Pallas will release his version soon, then we will know what he did.

Good to know, needing to use multiple wallets in general is very messy along with multiple proxies, multiple miner names or keys as well.

I'm not touching Pascal again till things get fixed up. Definitely need a pool or all these other issues need to be cleaned up for solo mining with a large enough farm. I can't just 'trust' things are working in the background and leave my miners on them for 12 hours only to find out some small detail is messing everything up. With pools you can actually see submitted shares, hashrate, and output... You're basically flying blind with solo mining unless the blocktime is low enough you can get blocks on a regular basis to make sure everything is working.

[mirny]

We need more solominers or a working stratum implementation.. . 3 hours without a single solo block.

Epsylon? Can you do it, and add pascalcoin to yiimp.

240 000 pascal coins are mined every day. ($180 000). If nanopool has 90% of the nethash, they earn $3240 everyday with the 2% fee.


Nanopool (AMD only) has most of the hashrate

where did you get those numbers?
afaik its 288blocks x 100coins/pb = 28800 in total for 24 hours

[bathrobehero]

Pascal's diff is so high that even a flat 3x speed increase would not make it the best choice to mine.

[Amph]

well it's dead already, in the sene that it's no different than mining all the other coins out there

waiting again for the next big thing...i missed it when no one was mining in actober, when the miner was launched

[bensam1231]

Pascal's diff is so high that even a flat 3x speed increase would not make it the best choice to mine.

According to cryptocompare, 3x speed would give you around $4 a card.

It's just not profitable to mine right now compared to other algos IF the pool and miners were all working properly. Pretty sure there is already work under way for private Pascal miners, if they don't already exist. Pascal has a a lot of available overhead. Each one of my rigs draws about half full TDP, so there is a lot of room there.

A lot of animosity surrounding this coin. Everyone is doom and gloom while we're waiting for devs to start ramping up the miners and iron things out. This still could be a very healthy market for Nvidia depending on where the cards end up after being optimized.

[pallas]

My version of Vorksholk/PascalCoin-CUDA, with fixes and speedups.

BTC donations: 1H7qC5uHuGX2d5s9Kuw3k7Wm7xMQzL16SN

https://github.com/pallas1/Pascalcoin-CUDA

Work in progress...

[7enuke]

My version of Vorksholk/PascalCoin-CUDA, with fixes and speedups.

BTC donations: 1H7qC5uHuGX2d5s9Kuw3k7Wm7xMQzL16SN

https://github.com/pallas1/Pascalcoin-CUDA

Work in progress...

windows binary ?

[ioglnx]

Hi Pallas,

could you please include the proxy Jar to have the full package.
I assume its not working against a pool or like the pascalopencl miner where u enter an ip and port :-D

[pallas]

Hi Pallas,

could you please include the proxy Jar to have the full package.
I assume its not working against a pool or like the pascalopencl miner where u enter an ip and port :-D

No pool, sorry.
And as I've said before, I can't test block generation (because my farm is too small), so it could be broken.

[ioglnx]

Me neither has a farm - just 2 Rigs with 12 Cards :-/
Even tho the output is 12GH/s i didn't make any block I think. Crap Nanopool and AMD.

[tbearhere]

Read and replied

You've got my skype and facebook name (check in my signature for confirmation eventually). Write me on fb, I'm on the move

I don't use face book