Nexus - Pure SHA3 + CPU/GPU + nPoS + 15 Active Innovations + More to Come

[kev7112001]

Code:

[METERS] 61 Block(s) ACC=60 REJ=1| Height = 24841 | Diff = 5.953534 | 35:02:16
  Prime Clusters Found: Four=75394 | Five=2890 | Six=63 | Seven=0 | Eight=0

  Not all of them.... This is the stat for my cc2.8x instance on Amazon.

Code:

[METERS] 66 Block(s) ACC=65 REJ=0| Height = 25888 | Diff = 6.059771 | 56:46:07
---------------------------------------------------------------------
ch      | 3        |  4        |  5        |  6        |  7        |
---------------------------------------------------------------------
count   | 1487476  |  87864    |  3029     |  71       |  1        |
---------------------------------------------------------------------
ch/m    | 436.707  |  25.796   |  0.889    |  0.021    |  0.000    |
---------------------------------------------------------------------
SPS 18.91 - Sieve/Test Ratio: 44.3 / 55.7 %

What is the spec of your rig?

3 780Ti cards, one core i7 CPU to go with it.



Your CPU miner looks heavily customized. Also you're into a bit of a gamble renting CPU power to mine CSD. After all the coin has to gain some value for you to recoup the costs.

Christian

if i were u i would move that psu u have it sucking in the hot air from the cards which can kill it and definitely lowers efficiency

[djm34]

thank you supercomputing, the miner compiles and works fine

        ***** skMiner for nVidia GPUs by djm34  *****
                     This is version v0.1
        based on ccMiner by Christian Buchner and Christian H. 2014 ***
                   and on primeminer by Videlicet
         Copyright 2014 djm34
          BTC donation address: 1NENYmxwZGHsKFmyjTc5WferTn5VTFb7Ze

GPU 1 = GeForce GTX 660

i wonder why my 660 with a small oc (100/200) is faster than your 750ti...  

The 660 has more cores than the 750ti... and this algo isn't doing lots of memory access...

[djm34]

Linux makefiles with some output enhancements for djm34's miner:

https://www.dropbox.com/s/1jth3nzfexg65t2/skminer.tar.gz?dl=0

tar -xvf skminer.tar.gz

cd skminer

./autogen.sh

./configure

make


Don't forget to change the NVCC options in the Makefile to match your GPU's compute capability:
-gencode=arch=compute_50,code=\"sm_50,compute_50\"

great you'll have more work soon...

Code:

[METERS] 61 Block(s) ACC=60 REJ=1| Height = 24841 | Diff = 5.953534 | 35:02:16
  Prime Clusters Found: Four=75394 | Five=2890 | Six=63 | Seven=0 | Eight=0

  Not all of them.... This is the stat for my cc2.8x instance on Amazon.

Code:

[METERS] 66 Block(s) ACC=65 REJ=0| Height = 25888 | Diff = 6.059771 | 56:46:07
---------------------------------------------------------------------
ch      | 3        |  4        |  5        |  6        |  7        |
---------------------------------------------------------------------
count   | 1487476  |  87864    |  3029     |  71       |  1        |
---------------------------------------------------------------------
ch/m    | 436.707  |  25.796   |  0.889    |  0.021    |  0.000    |
---------------------------------------------------------------------
SPS 18.91 - Sieve/Test Ratio: 44.3 / 55.7 %

What is the spec of your rig?

3 780Ti cards, one core i7 CPU to go with it.



Your CPU miner looks heavily customized. Also you're into a bit of a gamble renting CPU power to mine CSD. After all the coin has to gain some value for you to recoup the costs.

Christian

if i were u i would move that psu u have it sucking in the hot air from the cards which can kill it and definitely lowers efficiency

actually hot air from the card is expelled at the other end of the card...

[cbuchner1]

thank you supercomputing, the miner compiles and works fine

        ***** skMiner for nVidia GPUs by djm34  *****
                     This is version v0.1
        based on ccMiner by Christian Buchner and Christian H. 2014 ***
                   and on primeminer by Videlicet
         Copyright 2014 djm34
          BTC donation address: 1NENYmxwZGHsKFmyjTc5WferTn5VTFb7Ze

GPU 1 = GeForce GTX 660

i wonder why my 660 with a small oc (100/200) is faster than your 750ti...  

The 660 has more cores than the 750ti... and this algo isn't doing lots of memory access...

here's what me 3 780Ti's are doing.

98.6 MH/s | 24 Blks ACC=21 REJ=3 | Height = 27132 | Diff = 36 0-bits | 04:08:51

21 accepted blocks in 4 hours is great. I'll switch from the CPU channel to the GPU channel for a while.

[PVmining]

Linux makefiles with some output enhancements for djm34's miner:

big thanks supercomputing!
 
First i found it strange that the diff switched from 6.x to something like 0.00000091. Now I understand that the wallet switches from cpu channel to gpu channel... and there is a other diff. Right?

I have not found a single block the last hours (with more mhash than christian) - find it a little curious.
Strange... i tested yesterday the windows version of the miner a few minutes and found two blocks instantly.


EDIT: I think the miner makes something stupid. I get 35Mhash per ti. That seems clearly to much if I look at the numbers from christians 780ti or djm's 980. compiled the supercomputing source with compute 5.0

[djm34]

Linux makefiles with some output enhancements for djm34's miner:

big thanks!
 
First i found it strange that the diff switched from 6.x to something like 0.00000091. Now I understand that the wallet switches from cpu channel to gpu channel... and there is a other diff. Right?

I have not found a single block the last hours (with more mhash than christian) - find it a little curious.
Strange... i tested yesterday the windows version of the miner a few minutes and found two blocks instantly.

He is always lucky at finding blocks 
actually I turned to coinshield and got 2 blocks in less than 10 minutes...

[PVmining]

Linux makefiles with some output enhancements for djm34's miner:

big thanks!
 
First i found it strange that the diff switched from 6.x to something like 0.00000091. Now I understand that the wallet switches from cpu channel to gpu channel... and there is a other diff. Right?

I have not found a single block the last hours (with more mhash than christian) - find it a little curious.
Strange... i tested yesterday the windows version of the miner a few minutes and found two blocks instantly.

He is always lucky at finding blocks  
actually I turned to coinshield and got 2 blocks in less than 10 minutes...

I made a edit at the same moment above... do you think my hashrate could be right?
35mhash per 750ti - 210mhash per 6 card rig? I didn't think so...

[djm34]

Linux makefiles with some output enhancements for djm34's miner:

big thanks!
 
First i found it strange that the diff switched from 6.x to something like 0.00000091. Now I understand that the wallet switches from cpu channel to gpu channel... and there is a other diff. Right?

I have not found a single block the last hours (with more mhash than christian) - find it a little curious.
Strange... i tested yesterday the windows version of the miner a few minutes and found two blocks instantly.

He is always lucky at finding blocks  
actually I turned to coinshield and got 2 blocks in less than 10 minutes...

I made a edit at the same moment above... do you think my hashrate could be right?
35mhash per 750ti - 210mhash per 6 card rig? I didn't think so...

It isn't 35MH/s per 750ti (at least not mine) I think it is rather 26~28MH/s, also it could be some wallet issues, sometimes it gets stuck...

[antonio8]

Is there a link for the windows gpu miner?

[PVmining]

Is there a link for the windows gpu miner?

here my dear:
https://bitcointalk.org/index.php?topic=657601.msg9287828#msg9287828

[PVmining]

It isn't 35MH/s per 750ti (at least not mine) I think it is rather 26~28MH/s, also it could be some wallet issues, sometimes it gets stuck...

thanks for your answer.
re-checked  - with the test windows system I find a block quite fast, without restarting the wallet.
And the hashrate is clearly not the same as in linux (linux = 35mhash per card vs. 25mhash with windows).
...here seems something wrong. does your miner has a debug mode (-dp)?

[bigreddmachine]

Stuck at the block 15257 for hours WTF , what shall i do now

You should delete appdata, resync with add node in config file:
addnode=69.195.149.114
addnode=204.27.62.234
addnode=69.195.148.42
addnode=204.27.62.242
addnode=69.195.148.34
addnode=204.27.62.226
addnode=62.219.98.129

After syncing, diff should be 6.xx

My wallet is showing only 1 network connection ATM.  Are these nodes still viable?  Should I add them to my config file?

[djm34]

It isn't 35MH/s per 750ti (at least not mine) I think it is rather 26~28MH/s, also it could be some wallet issues, sometimes it gets stuck...

thanks for your answer.
re-checked  - with the test windows system I find a block quite fast, without restarting the wallet.
And the hashrate is clearly not the same as in linux (linux = 35mhash per card vs. 25mhash with windows).
...here seems something wrong. does your miner has a debug mode (-dp)?

nope... may-be you could try to run on the testnet and see what happens...
had a rapid look into the main loop, the linux miner use a different method to access time, would that be the problem ? (don't know enough about these routines...).
But I don't know if it is really related, looking at Christian hashrate actually it should be higher (on windows the 780ti does around 40MH/s...)

[antonio8]

Is there a link for the windows gpu miner?

here my dear:
https://bitcointalk.org/index.php?topic=657601.msg9287828#msg9287828

Thanks PV.

I guess  I should have used that search button.

[blaaaaacksuit]

Anyone want to buy my coinshield?  1000 for 0.4 BTC. PM me if you're interested.

[djm34]

Anyone want to buy my coinshield?  1000 for 0.4 BTC. PM me if you're interested.

I can sell you mine at this rate if you want   (at that price, I will start renting aws session)

[blaaaaacksuit]

Anyone want to buy my coinshield?  1000 for 0.4 BTC. PM me if you're interested.

I can sell you mine at this rate if you want   (at that price, I will start renting aws session)

LOL, didn't notice they were up on an exchange finally.  Been away for a while. 

[blaaaaacksuit]

Did the developer ever make the source code public for this?

[skunk]

Did the developer ever make the source code public for this?

not yet

@supercomputing: are you still finding blocks under linux? still none here after ~8h...

[PVmining]

nope... may-be you could try to run on the testnet and see what happens...
had a rapid look into the main loop, the linux miner use a different method to access time, would that be the problem ? (don't know enough about these routines...).
But I don't know if it is really related, looking at Christian hashrate actually it should be higher (on windows the 780ti does around 40MH/s...)

good idea. I will try the testnet - but I'm pretty sure that something in the linux version is broken (on my rigs).
with 1/3 of my hashpower i found with the windows version in 2 hours 4 blocks.
with the linux version nothing - all rigs combined in a timeframe of six hours.

@supercomputing: are you still finding blocks under linux? still none here after ~8h...

me too - if you read my short novel above...