[ANN] ccminer 2.3 - opensource - GPL (tpruvot)

[sp_]

It needs complete rewrite. simd512 will be the most difficult

This code is not based on the 4way groestl in ccminer, but the opencl wolf0 groestl kernel in sgminer. I have updated the post with the more code.

[santan]

KRLX can earn 0.25-0.5 BTC in bounties if he publish his work.

They want windows version. I dont have windows computers.

Your skunk miner was so easy to compile in windows. so many users here did it, including me.
You will find here hundreds of sincere guys who can do that for you.

[halker2010]

KRLX can earn 0.25-0.5 BTC in bounties if he publish his work.

They want windows version. I dont have windows computers.

Your skunk miner was so easy to compile in windows. so many users here did it, including me.
You will find here hundreds of sincere guys who can do that for you.

correct , please release it haters gonna hate and whiners gonna whine anyway.

[xs.over]

I see a lot of people are looking for xevan algorithm to work at ccminer, and some developers sell ccminer copy for 0.5BTC.

Does these people know what xevan algo coins have very shall hashrate? 200+ pcs gpus like gtx 1070 and profit almost gone

[antantti]

You want some profitable nvidia mining to happen? Next time something new is released contact tpruvot.

[bensam1231]

There are always people willing to compile something for Windows if you post it. He's currently getting his bounty out of Xevan and then can release it for the .5 bounty. Either way, once it's released it wont earn more then all the other algos at the moment. Too public, too popular, the new kids still don't know how difficulty works. Only reason it looks more profitable (and not by that much) is because one person is currently mining it.

Krlnx knows this and that's why he publicly announced it. It's no longer that profitable and he probably can scoop up quite a bit more from private offers in PMs.

You want some profitable nvidia mining to happen? Next time something new is released contact tpruvot.

Epsylon doesn't make ground up miners AFAIK. I've tried contacting multiple developers to get private miners made and most wont even talk about it despite willing to compensate them quite well. It's actually pretty ludicrous. You'd think Cryptos would have attracted a lot more talent then they have as far as programming goes.

[antantti]

You'd think Cryptos would have attracted a lot more talent then they have as far as programming goes.

Someone pays more. No matter how big farm you build some builds bigger.

I'm sure you have seen this, could be fake:

[bensam1231]

You'd think Cryptos would have attracted a lot more talent then they have as far as programming goes.

Someone pays more. No matter how big farm you build some builds bigger.

I'm sure you have seen this, could be fake:

Not sure what this means? There aren't developers because there is someone willing to pay more? You don't even know what 'more' is in this situation.

[antantti]

Not sure what this means? There aren't developers because there is someone willing to pay more? You don't even know what 'more' is in this situation.

Sorry, some editing, stupid me.

Let me put it this way: You are not paying enough.

[bensam1231]

Not sure what this means? There aren't developers because there is someone willing to pay more? You don't even know what 'more' is in this situation.

Sorry, some editing, stupid me.

Let me put it this way: You are not paying enough.

You don't know how much I'm paying. Despite leaving open ended amounts to those I wish to employ they don't even look to do business with you.

[santan]

Actually krnlx is doing it right. Nobody pays a penny if he release the miner in public now. So it's better that he is mining himself, maybe only with few GPU's, to get his hard work pay for himself.
Everyone just wait and not destroy his profit.

[blackhorse7]

Mister Pruvot, I have a request for ccminer:please add a second benchmark mode -precise when I not use -a switch with benchmark and miner makes benchmarks for all algirithms in a bunch

Code:

ccminer-x64 --benchmark --no-color > ccminer-x64.txt

Code:

[2017-10-01 08:00:57] Benchmark results for GPU #0 - EVGA GTX 1060 6GB:
[2017-10-01 08:00:57]    blakecoin :    2980887.7 kH/s, -5955 MB, 67108864 thr.
[2017-10-01 08:00:57]        blake :    1704596.5 kH/s,     2 MB, 67108864 thr.
[2017-10-01 08:00:57]      blake2s :    2610601.3 kH/s,     2 MB, 33554432 thr.
[2017-10-01 08:00:57]          bmw :    1183760.6 kH/s,    66 MB,  2097152 thr.
[2017-10-01 08:00:57]      bastion :       8165.2 kH/s,   234 MB,  1048576 thr.
[2017-10-01 08:00:57]         deep :      26029.9 kH/s,    34 MB,   524288 thr.
[2017-10-01 08:00:57]       decred :    1773028.7 kH/s,     2 MB, 33554432 thr.

For now benchmark mode is making a few seconds measurements and some are inaccurate (for example tribus benchark shows 20Mh, but with switch -a tribus it shows 40Mh). With this mode miner could in addition make power measurement for each algorithm during the test - we know, that some of algorithms are less power hungry and that measurement could help to better calculate profit. Now with those 2-5 seconds measurements when I try to observe power on GPU-Z it is imposible because when GPU-Z shows power, miner is going to switch to another algo.
That change should not be hard to do, but it should make benchmarking a lot easier and accurate.
My vision:

Code:

[2017-10-01 08:00:57] Benchmark results for GPU #0 - EVGA GTX 1060 6GB:
[2017-10-01 08:00:57]    blakecoin :    2980887.7 kH/s, -5955 MB, 90W, 67108864 thr.
[2017-10-01 08:00:57]        blake :    1704596.5 kH/s,     2 MB, 90W, 67108864 thr.
[2017-10-01 08:00:57]      blake2s :    2610601.3 kH/s,     2 MB, 95W, 33554432 thr.
[2017-10-01 08:00:57]          bmw :    1183760.6 kH/s,    66 MB, 93W,  2097152 thr.
[2017-10-01 08:00:57]      bastion :       8165.2 kH/s,   234 MB, 90W,  1048576 thr.
[2017-10-01 08:00:57]         deep :      26029.9 kH/s,    34 MB, 93W,   524288 thr.
[2017-10-01 08:00:57]       decred :    1773028.7 kH/s,     2 MB, 69W,  33554432 thr.

[antantti]

You don't know how much I'm paying. Despite leaving open ended amounts to those I wish to employ they don't even look to do business with you.

True, I don't. I guess it's safe to say money isn't the reason for your recruiting problems, it must be something else...

[seasonw]

I see a lot of people are looking for xevan algorithm to work at ccminer, and some developers sell ccminer copy for 0.5BTC.

Does these people know what xevan algo coins have very shall hashrate? 200+ pcs gpus like gtx 1070 and profit almost gone

I don't quite get you... Actually any algo with large miners, surely the profit will be lesser...

[Epsylon3]

if you want to talk about xevan, go in BSD or XLR threads... i already answered about that

[santan]

Epsylon3,
If i need to start from a random extranonce2.......... should i just change in this line in ccminer.cpp

Code:

for (i = 0; i < (int)sctx->xnonce2_size && !++sctx->job.xnonce2[i]; i++);

as    i = rand();   
or any other file change needed ?

[LuxLP]

Hi,
Api isn't workng in my system, although I specify api bind as a current ip in the local network, set allow remote and api allow as 0/0. Ebfw works fine in the same system.
Thanks!

[joblo]

Epsylon3,
If i need to start from a random extranonce2.......... should i just change in this line in ccminer.cpp

Code:

for (i = 0; i < (int)sctx->xnonce2_size && !++sctx->job.xnonce2[i]; i++);

as    i = rand();   
or any other file change needed ?

Why must it be random? Regardless, your proposed change is way off base.
Take a look at util.cpp:stratum_notify for the start value.

[fr4nkthetank]

< the new kids still don't know how difficulty works >

lol that is true though, they got farms of nvidia cards, mining these small coins...and for what ?  a bit more profit than zec maybe.  nah, those smaller coins, you control the supply, corner the coin, and pump it  but hey what do i know i dont have a lambo

[sp_]

bitcore trouble @ yiimp again...