[ANN] lolMiner-mnx, Mining MinexCoin(MNX) New version: v0.4 (July 22)

[Lolliedieb]

lolMiner-mnx
-- Announcement Thread --

Hi folks,
this is the thread for announcing lolMiner-mnx for mining MinexCoin (MNX) based on the Equihash 96/5 algorithm (please do not confuse this with Zcash / BitcoinGolds Equihash 200/9).

The new versions from 0.4 onwards got a new thread, because I added there other coins as well.
See here: https://bitcointalk.org/index.php?topic=4724735

[AAile]

reserved

[Dirtpaw]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

[Kompik]

Thanks for the miner, but i dont really get how to start a 6 gpu rig Your comment with platform is confusing, i dont understand it at all

[salshicha]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

Can you tell me where to find OpenCL miner for windows?

[Lolliedieb]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

Well technically OpenCL is not necessarily slower than CUDA. Usually is is only that most OpenCL codes are optimized for AMD and the OpenCL bindings of AMD are a bit more efficient. But generally I think improving the algorithms behind will bring more sol/s than changing to CUDA which would be a pain for me to learn

By the way: my development and testing machines have Nvidia GPUs running under Linux. So the current version is rather optimized for NVidia. For AMD I only know that the miner works, but I am completely in the dark how fast it is. So at this time this is a Nvidia optimized software.

By the way: I will surely lower the fee to 1% in the future. But right now I just have to get a start

Thanks for the miner, but i dont really get how to start a 6 gpu rig Your comment with platform is confusing, i dont understand it at all

Well you can have multiple drivers on your system, e.g. when you have graphic cards from multiple vendors in your rig. Then the OpenCL drivers are organized in platforms, so all cards belonging to the one vendor are in platform 0, all from the second in platform 1 and so on.

As for the device selection the devices are sorted by their platform.
Lets assume that you have all 6 GPUs from the same GPU Chip Vendor (AMD or Nvidia ... the concrete creator of the card (Gainward, MSI, ... ) do not matter). Then it is likely that you have only one platform on your machine that has index 0.
So you then can start the miner just 6 times, but with different device parameter:
--device 0
--device 1
...
--device 5

Then all cards should get utilized well.

[AAile]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

Can you tell me where to find OpenCL miner for windows?

mxMiner download

[salshicha]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

Can you tell me where to find OpenCL miner for windows?

mxMiner download

I read about, but is it faster than lolminer?

[Aeon_]

Tested under nvOc 19.4

Card: MSI Gtx 1070

Settings:
INDIVIDUAL_POWERLIMIT_0=110
__CORE_OVERCLOCK_0=150
MEMORY_OVERCLOCK_0=800

Hash: 14.5Kh/s

Good job;D

[AAile]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

Can you tell me where to find OpenCL miner for windows?

mxMiner download

I read about, but is it faster than lolminer?

I'll test it soon.
So far, virustotal says the file is 100% virus free.

[Dirtpaw]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

Well technically OpenCL is not necessarily slower than CUDA. Usually is is only that most OpenCL codes are optimized for AMD and the OpenCL bindings of AMD are a bit more efficient. But generally I think improving the algorithms behind will bring more sol/s than changing to CUDA which would be a pain for me to learn

By the way: my development and testing machines have Nvidia GPUs running under Linux. So the current version is rather optimized for NVidia. For AMD I only know that the miner works, but I am completely in the dark how fast it is. So at this time this is a Nvidia optimized software.

By the way: I will surely lower the fee to 1% in the future. But right now I just have to get a start

Thanks for the miner, but i dont really get how to start a 6 gpu rig Your comment with platform is confusing, i dont understand it at all

Well you can have multiple drivers on your system, e.g. when you have graphic cards from multiple vendors in your rig. Then the OpenCL drivers are organized in platforms, so all cards belonging to the one vendor are in platform 0, all from the second in platform 1 and so on.

As for the device selection the devices are sorted by their platform.
Lets assume that you have all 6 GPUs from the same GPU Chip Vendor (AMD or Nvidia ... the concrete creator of the card (Gainward, MSI, ... ) do not matter). Then it is likely that you have only one platform on your machine that has index 0.
So you then can start the miner just 6 times, but with different device parameter:
--device 0
--device 1
...
--device 5

Then all cards should get utilized well.

Good to hear. I will test out the miner and check how much faster it is, if it is

Edit: Tested the miner. If i run only 1 gpu then the speed is almost double of optiminer but when I run 6 GPUs hashrate drops quite alot and is slower than Optiminer. Also one GPU dropped core when I started miner (maybe picky gpu doesnt like same clocks as other cards.). Tested with 1070 ti-s.

[Kompik]

Dev-fee is too high and we are missing CUDA optimized miner. We already have OpenCL miner. So I suggest you lower the dev-fee and make a CUDA version.

Well technically OpenCL is not necessarily slower than CUDA. Usually is is only that most OpenCL codes are optimized for AMD and the OpenCL bindings of AMD are a bit more efficient. But generally I think improving the algorithms behind will bring more sol/s than changing to CUDA which would be a pain for me to learn

By the way: my development and testing machines have Nvidia GPUs running under Linux. So the current version is rather optimized for NVidia. For AMD I only know that the miner works, but I am completely in the dark how fast it is. So at this time this is a Nvidia optimized software.

By the way: I will surely lower the fee to 1% in the future. But right now I just have to get a start

Thanks for the miner, but i dont really get how to start a 6 gpu rig Your comment with platform is confusing, i dont understand it at all

Well you can have multiple drivers on your system, e.g. when you have graphic cards from multiple vendors in your rig. Then the OpenCL drivers are organized in platforms, so all cards belonging to the one vendor are in platform 0, all from the second in platform 1 and so on.

As for the device selection the devices are sorted by their platform.
Lets assume that you have all 6 GPUs from the same GPU Chip Vendor (AMD or Nvidia ... the concrete creator of the card (Gainward, MSI, ... ) do not matter). Then it is likely that you have only one platform on your machine that has index 0.
So you then can start the miner just 6 times, but with different device parameter:
--device 0
--device 1
...
--device 5

Then all cards should get utilized well.

Thanks will try it out. I am also getting lilboost system md error on some rigs?

[salshicha]

When i use a few instances, some of them crash and close

[mk111]

How many sols does a GTX 1070 and a RX 580 do on the Windows version?

[Lolliedieb]

Thanks will try it out. I am also getting lilboost system md error on some rigs?

Thats interesting... on some but not on all? Can you provide more info about your configuration and how the rigs may differ? From libbost system md I assume that the rigs are running Windows?

When i use a few instances, some of them crash and close

The maximum that I tested here were 2 instances running on two GTX 1080... So I do not know yet what happens with more. Can ypu give me more info about your config and what error message is given at crash time? Hope we can find the issue quickly.

 
I will think during this evening how I can thank you early testers reporting bugs and hash rates. Will give a post on that tomorrow

[salshicha]

Thanks will try it out. I am also getting lilboost system md error on some rigs?

Thats interesting... on some but not on all? Can you provide more info about your configuration and how the rigs may differ? From libbost system md I assume that the rigs are running Windows?

When i use a few instances, some of them crash and close

The maximum that I tested here were 2 instances running on two GTX 1080... So I do not know yet what happens with more. Can ypu give me more info about your config and what error message is given at crash time? Hope we can find the issue quickly.

 
I will think during this evening how I can thank you early testers reporting bugs and hash rates. Will give a post on that tomorrow

Tried 5 instances with nvidia 1060 and W10

2 of them eventually crashed with windows message "app stopped working"

[AAile]

First of all, wow, lolMiner-mnx gives great hashrates, almost 10 times better than mxMiner_0.20.1.
One gtx 1060 mini gives around 8000 KSol/s.

The issues:
It keeps the CPU at 99% busy all the time, even if you don't use cpu mining.
It does crash randomly in about 15 minutes of mining. Doesn't look related to anything, could be an overflow or some other leak in the code.
I tested this on my test rig with only 2 cards (gtx 1060 gtx) with two threads for each card. I have tested the cpu mining shortly, but without the cards.

You're on a good path to create a very good miner.

[Aeon_]

How many sols does a GTX 1070 and a RX 580 do on the Windows version?

GTX 1070
Win10
@+0 cc /+800 mem
Average speed (5s): 14444.7 Sol/s
Average speed (5s): 14136.5 Sol/s
Average speed (5s): 14530.8 Sol/s
Average speed (5s): 14531.3 Sol/s

Edit: CPU i5 3570k 78% utilization with two instances

Edit2: stable for 15-20 min then crashed

[salshicha]

With an i5-4690k , more than 3 instances of the miner with 1060s, the hashrate of every instance falls dramatically

When the miner crases, the error relates to libstdc++-6.dll


Can you put a parameter on miner to set -cl-local-work?

Perhaps the solution another miner developer adopted may be useful to you

He made a opencl miner that eat all cpu available, and the more gpus working, less performance for each one

There is an issue with NVIDIA GPUs leading to very high CPU usage. The reason is improper implementation of OpenCL by NVIDIA. When CPU thread waits for results from GPU, it does not stop, it spins in loop eating CPU resources for nothing.  
 There was impemented a workaround on this issue: before reading results from GPU current thread sleeps during small calculated time.

He resolved it and the miner is more stable and a lot faster.

[qwep1]