Topic: [ANN] lolMiner-mnx, Mining MinexCoin(MNX) New version: v0.4 (July 22) - page 43. (Read 30772 times)

TeamViewer generally eats graphic power, see here:
https://www.reddit.com/r/gpumining/comments/7mwgpc/high_gpu_usage_in_teamviewer_pc_mining/

@All that have problems with the --enable-nv-hotfix:
It seems that on some systems the current version is to aggressive. I will try an alternative variant in the next hours and hopefully find a solution that works for all users. If new one looks more promising, will be included in the next version.


PS: I currently had access to a Linux rig with 8 1080's all stock. Miner runs now for about 16 hours stable there without any crash. It happens that one of the instances looses connection, but reconnect worked fine for this ones.

Running fairly stable here on W10, except that when I shut off teamviewer, hash rate drops from 14ksols to 1500 sols...
Then jumps right back up when team viewer is turned on again.

Ideas?  

Edit: Removing hot fix solved speed issue, but now miner keeps crashing 

First of all I want to thank you for creating this miner, and I know myself how much effort has to be put into software development, so I don't think 2% dev fee is a problem, especially if you get 10 times the sols you got before.

I tested it in a couple of different environments, all windows for the moment, linux will follow, all cards with stock settings:
1050ti => 5.5ksol
980 => 10ksol
290x => 7k
580 => 6.5ksol

tomorrow i should get 2 vega rx 56 will test them too then.

on one win10 pc i have the following:
Platform 0: NVIDIA CUDA, OpenCL 1.2 CUDA 9.1.83
   Device 0: GeForce GTX 980
Platform 1: AMD Accelerated Parallel Processing, OpenCL 2.0 AMD-APP (2442.12)
   Device 0: Hawaii
   Device 1: Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz
Platform 2: Intel(R) OpenCL, OpenCL 2.0
   Device 0: Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz

trying tp run cpu under platform 1 make the miner crash immediately, under platform 2 it works however.

Not worthless at all: you can use a shell script for an automatic relaunch:
 in a .bat:

echo -------------------------------------------------------------------
echo Simple script to restart your miner software after a period of time
echo -------------------------------------------------------------------
echo:

set executable=lolMiner-mnx.exe
set commandline= --server mnx.suprnova.cc --port 7076 --user xxxxx  --pass x    --platform 1
set /a counter=0
:start
%executable% %commandline%
echo:
echo Restarting the software  (%counter%)
set /a counter+=1
echo:
echo:
goto start

I have similar behaviour on a titan xp. Far worse performance with --enable-nv-hotfix.

System : windows 10 (creator update)
cpu : xeon E5 2696V3 (18 cores @3.2ghz with microcode/bios hack).
With the hotfix enables, performance is aroung 5 instead of 22-23. Cpu usage is very low ( less thant 20% for a dedicated core). Without the hotfix, performance is good and cpu usage is near a full dedicated core.

1080ti aorus extreme (×1 for testing)
win10
Current Cuda 9
With and without --enable-nv-hotfix
Crashes within 10-15 min

Without Hotfix 20-23 ksol
With Hotfix 5-6 ksol

Done without any adjustments

=> worthless cause u cant leave it unattended

Linux not tested Till now.

I've used lolminer with one gpu on windows and it seems to crash frequently. Every couple of hours, I have to reset the miner. Hope to see updates to fix the problems.

AMD athlon II X2 2.9GHz - very simple and low cost

But what is your cpu, even if its HyperThreaded, use the number of physical cores it has. otherwise crashes or slowdowns happen. 1 core per card.

On windows 10 x64 (without --enable-nv-hotfix)
1x1070 - 14.0-15.0
2x1070 - 10.0-10.5
3x1070 - 7.8-8.2
1x1050 - 5.0-5.2

On windows 10 x64 (with --enable-nv-hotfix)
1x1070 - 1.8-3.0
2x1070 - 1.8-3.0
3x1070 - 1.8-3.0
1x1050 - 5.0-5.2

On ubuntu 16.04 (with --enable-nv-hotfix)
3x1070 - 13.2-14.0 (core 150, mem 300)
4x1060 - 8.7-9.2 (core 100, mem 200)
5x1050ti - 5.3-5.8 (core 50, mem 0)


Best solution for linux, more hashes than optiminer
Uptime on linux 8 hours

10 minutes work - error. I want new version!

See two posts above - will be fixed in next version since I now know what is the problem


Not that valuable comment  Can you be more specific? 

Error error error error error
 

I came across a rather weird incident. Suddenly all instances dropped connection to stratum and started reconnecting. After 20 min of reconnecting 4 gpus out of 6 reconnected but other 2 just keep reconnecting. Any solutions? Restarting those instances does not help. Running on Linux.

Update: other 2 reconnected 10 min after the others.

Hmm well that one is new. Maybe - I think Win 7 tells us more then Win 10 does.
That libstdc++ is the standard library required for any c++ build with gcc compilers. I should check if there are recent bug-fixed versions of this library that also operate with the miner binary. Thanks for reporting.

By the way I think that I found at least one of the issues preventing the miner to start on some AMD platforms (that build error). Was a glitch in OpenCL code and will be fixed in next version

Also I identified two possible bottlenecks why Vega scales well (but not so well as GTX 1080 and up) and why the Polaris cards are slower then their Nvidia competitors (we all expect an RX 580 to be faster then a 1060, but currently the 1060 seems to be 3-4k sols ahead. It is interesting to see how much things depend on few optimizations for the one or the other side. Currently for AMD (with Linux) I would prefer optiminer / 1.2 , since it is faster, but lolMiner is for Nvidia cards. Strange isn't is? In particular since both use OpenCl...

Anyways - I now have an idea but I am not certain how easy it is to fix without lowering the good performance on Nvidia. Currently I would like to focus on stability and usability before doing too much kernel work to boost up AMD miners. When most crashes are fixed I of cause will come back to this topic. In the meantime optiminer surely still does its job (And I gain some time getting me a Windows / AMD system for debugging purposes ^^)

By the way thanks @Syth for posting the virus check results. 

here's a virus scan of the latest version (0.11)
https://www.virustotal.com/#/file/86ee695c995fa3027505a7df1f114c3c22f59bd7af16d5026277358bf07a9c85/relations

I do envy you as running more than two instances would result in decreased hash rate on any of my rigs with 1070, 1070ti or 1080ti... I think I  will have to wait for the next release.

Tried it on my HTPC (Win7, GT1030).
For the first 10 attemtps miner crashed after 10-30 seconds with Access Violation Error in libstdc++-6.dll.
I think it happend when submitting shares.
Now it's running stable since about 45 minutes with avg 2.15 KSol/s.

Nice job!

It started and seems like it's working. I'm getting Way more Sol than mxminer so Kudos to you sir! I hope you make better improvements to this miner and people will recognize it!

Working great for me on my rigs. Wondering if you will go a cuda route and see if the hashrate can be pushed further. If you do I can do some testing.