Author

Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 436. (Read 2347659 times)

newbie
Activity: 39
Merit: 0
I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?

That'd be great if it consistently gets that. The other miners also have extreme issues with CPU utilization. Only worth it if you get it working with pools. You can't solo mine anymore.

the hashrate is consistent and cpu usage is much lower.
the 970 does 500 Mh/s @100W.
the miner can't be easily adapted to pools, it's too basic.
Would you be willing to open source what you have now? I would love to help port it to pools if possible.
legendary
Activity: 1764
Merit: 1024
Just as a FYI, BoolBerry is still well worth looking into. Been asking for like six months. It has very healthy markets and hasn't really been improved on in a long time. Pretty sure there are private miners out there for that algo and coin.

BTW Pallas Nanopool has their own miner with pool support and source you can tear apart.
sr. member
Activity: 574
Merit: 250
Fighting mob law and inquisition in this forum
You answered to something where you even don't know if he used version of KlausT or other with pool support.
KlausT is very active :-D few minutes ago a commit..Klaus/Schleicher let me know where to send the Austrian beer :-D
hero member
Activity: 677
Merit: 500
the initial miner of volksheit didnt split the shares across the gpu. ever gpu got same units and the proxy just avoid duplicated to prevent ban. this is what i understood and see fromdebug logs of the miner.he also mention like that.

Yes it is broken in many ways: nonce ranges, kernel launch parameters, etc.
I may have fixed all that but I have not enough power to mine a block and, thus, test it fully.

Good work.

Can you please share you miner.
I don't know how many gpus you have I have currently 12 GPUs running..most of them 1080 :-D
Solo is dead... All blocks is own by nanopool... Sad
sr. member
Activity: 574
Merit: 250
Fighting mob law and inquisition in this forum
the initial miner of volksheit didnt split the shares across the gpu. ever gpu got same units and the proxy just avoid duplicated to prevent ban. this is what i understood and see fromdebug logs of the miner.he also mention like that.

Yes it is broken in many ways: nonce ranges, kernel launch parameters, etc.
I may have fixed all that but I have not enough power to mine a block and, thus, test it fully.

Good work.

Can you please share you miner.
I don't know how many gpus you have I have currently 12 GPUs running..most of them 1080 :-D
legendary
Activity: 2716
Merit: 1094
Black Belt Developer
the initial miner of volksheit didnt split the shares across the gpu. ever gpu got same units and the proxy just avoid duplicated to prevent ban. this is what i understood and see fromdebug logs of the miner.he also mention like that.

Yes it is broken in many ways: nonce ranges, kernel launch parameters, etc.
I may have fixed all that but I have not enough power to mine a block and, thus, test it fully.
legendary
Activity: 1176
Merit: 1015
When pool hashrate is much higher than network, it may be due to a high number of orphans, so be careful when you do your own math.

That could be the case here, only 25% of shares my hawaii's send to pool are valid.

I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?

If zec looses 80% of its current value there are many miners looking for a new coin to mine. Pasc would help nvidia owners a lot, amd cards are currently on 400-600MH range so 900MH would be huge advantage. Definitely worth working.
sr. member
Activity: 574
Merit: 250
Fighting mob law and inquisition in this forum
the initial miner of volksheit didnt split the shares across the gpu. ever gpu got same units and the proxy just avoid duplicated to prevent ban. this is what i understood and see fromdebug logs of the miner.he also mention like that.
legendary
Activity: 1134
Merit: 1001
None miner nvidia work on a pool .
legendary
Activity: 2002
Merit: 1051
ICO? Not even once.
I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?

That'd be great if it consistently gets that. The other miners also have extreme issues with CPU utilization. Only worth it if you get it working with pools. You can't solo mine anymore.

the hashrate is consistent and cpu usage is much lower.
the 970 does 500 Mh/s @100W.
the miner can't be easily adapted to pools, it's too basic.

I'm interested, let me know if I can help with testing.
sp_
legendary
Activity: 2954
Merit: 1087
Team Black developer
I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?

That'd be great if it consistently gets that. The other miners also have extreme issues with CPU utilization. Only worth it if you get it working with pools. You can't solo mine anymore.

the hashrate is consistent and cpu usage is much lower.
the 970 does 500 Mh/s @100W.
the miner can't be easily adapted to pools, it's too basic.

Did you add cpu verification?
legendary
Activity: 2716
Merit: 1094
Black Belt Developer
I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?

That'd be great if it consistently gets that. The other miners also have extreme issues with CPU utilization. Only worth it if you get it working with pools. You can't solo mine anymore.

the hashrate is consistent and cpu usage is much lower.
the 970 does 500 Mh/s @100W.
the miner can't be easily adapted to pools, it's too basic.
legendary
Activity: 1764
Merit: 1024
I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?

That'd be great if it consistently gets that. The other miners also have extreme issues with CPU utilization. Only worth it if you get it working with pools. You can't solo mine anymore.

This is sort of a weird predicament, I think a lot of people are put off by this coin (myself included) due to the shitty hodge podge of job they did on the wallet and the miners. Everything is held together by band-aids and isn't explained very well. I wasted 12 hours not knowing my miners were duplicating about 70% of their work because I didn't read the fine print (I'm sure I'm not the only one based on the confusion in this thread as well).

That being said, Pascal is an extremely healthy market for mining right now, which means if it can be optimized, it's in demand. It doesn't matter if this coin is a pump and dump or whatever, compared to other altcoins it's still a very green pasture. The hashrate battle is going to start now, which means optimizations are needed to make things more competitive (especially with AMD).

When the pools get ironed out and the miners get ironed out, this is going to be another core market even if it drops down to Lbry levels, it'll still be worth mining and having as a option.

legendary
Activity: 2002
Merit: 1051
ICO? Not even once.
Not sure why you guys keep quoting payment estimations. You can just type that into cryptocompare and gives you the same results, once again depending on nethash which is all over the fucking place and doesn't seem to be a accurate place for it yet (which is also why whattomines calculator isn't up yet).

Because more ways of estimating earnings gives us a better overall picture. And because the difficulty is the most reliable factor, not nethashrate.

It's not different, it's the same thing, the only thing that's different is the nethash you put into the equation.

Both nethashrate and the difficulty fluctuates a lot so we don't have precise estimates but they're very different metrics.

The difficulty is the clear and precise number of how much work has to be done to solve a block while nethashrate only shows us the total mining power on the network. If the difficulty retarget would be perfect, the two would be the same (as to showing profitability) but it's far from it. If the nethashrate is higher than what the difficulty should be to keep the blocktime then you just get more blocks mined than the intended blocktime and vice versa - which is why we had rapid blocks and very slow blocks.
sr. member
Activity: 574
Merit: 250
Fighting mob law and inquisition in this forum
I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?

Give it a shot pallas.
legendary
Activity: 2716
Merit: 1094
Black Belt Developer
I modified the pascal cuda miner and optimized it, now it's doing 900 Mh/s on a 1070 consuming just 100W.
Is it good? Worth working on it, now that there are pools?
sr. member
Activity: 574
Merit: 250
Fighting mob law and inquisition in this forum
KlausT was active and implemented already pascalcoin into his CCminer branch tonight with primitive stratum support.
But I already tried the build and it's doesn't seem to work against nanopools pascalcoin pool
That's because the coding is not finished yet.
For example all the CUDA code for pascal is still missing.


Klaus no worries I know it is not complete yet :-D It was no intention to say you made bad work or such.
If you felt hurt I'm sorry - I was just telling that you are on it and first attempt (der Hüftschuss) didn't worked for me :-D

Keep on good work.

Gruß^^
legendary
Activity: 1764
Merit: 1024
Not sure why you guys keep quoting payment estimations. You can just type that into cryptocompare and gives you the same results, once again depending on nethash which is all over the fucking place and doesn't seem to be a accurate place for it yet (which is also why whattomines calculator isn't up yet).

Because more ways of estimating earnings gives us a better overall picture. And because the difficulty is the most reliable factor, not nethashrate.

It's not different, it's the same thing, the only thing that's different is the nethash you put into the equation.
hero member
Activity: 675
Merit: 514
KlausT was active and implemented already pascalcoin into his CCminer branch tonight with primitive stratum support.
But I already tried the build and it's doesn't seem to work against nanopools pascalcoin pool
That's because the coding is not finished yet.
For example all the CUDA code for pascal is still missing.
legendary
Activity: 3248
Merit: 1072
this was just another coin to mine early and dump at 50-100k range or even 30k, huge profit
Jump to: