Topic: [ANN] lolMiner-mnx, Mining MinexCoin(MNX) New version: v0.4 (July 22) - page 46. (Read 30772 times)

thanks for the answer, but so far nothing has helped. I turn on the second bat file, it divides the power of one video card in half

One bat per card, selecting the right gpu

But it crashes often with more than 1 instance running

I have 5 GTX 1060 but I only get one card. you can explain in detail how to run all 5 cards

Win 10 with GTX 1070
Getting 14000 sols and been running stable for 2 hours.

When I start another instance to mine on a second GTX 1070 the sols drops on both cards and the second instance crashes after 5 minutes.

On windows 10, GTX1050 core 100, mem 350, 5.2-5.4ksol, uptime 4 hours, no errors (better then optiminer on linux)

Getting build (-11) error on tahiti cards (280x).. Please fix this in next version.

Btw geiler Nickname Lolliedieb     

I just tested on Linux (ethOS) RX560 - works stable but with low hash rate 2600 H\s vs optiminer 5000 H\s

Thank you very much for your work, your work ethic (transparency) and user support.

I can fully understand that - when you want to make money with mining stability is the alpha and omega. Thats why I released this so early and with hints that this is beta because I do not have the testing machines (yet) to get it rock stable for all platforms that it may support. Therefore I am very thankful about any comment here helping me to improve stability.

I will add an extra mark to the Windows version that it is instable. Linux version currently causes much less problems and should be preferred when possible.

Due to the feedback of last hours, here is the updated list of known issues:

[ALL OS, NVIDIA] High CPU Utilization (Found by me, Confirmed by all NVidia Users here)
Reason:Nvidia has a busy wait for OpenCL kernel completion thats considered a but for over 6 years now, but is not yet fixed (in CUDA it is...)
Problem for Miners: High CPU utilization that may cause problems with CPU cooling. Also dropping hashrate when too many instances are active on the same rig with week CPU and instances may crash. For example an Intel Celeron 3855u may only run one instance of lolMiner currently, Core I5 with 4 physical cores may run up to 3 GPU. More then 4 GPU are currently only stable with Intel Server CPU or AMD Threadripper.
Expected Time to Fix: I am currently testing a fix that will lower the CPU cost of one GTX 1080 from about 115% single core CPU utilization to about 20 (on my testing machine). This will cost about 2% of hash-rate but will allow larger rigs to operate and the CPU to stay cooler. I will provide the beta fix today before 3pm UTC.

Ps: You will always require a strong CPU for Minexcoin anyways, since the found solutions have to be put twice to SHA256 to compare it with the target. This is currently done on CPU and since the high end Nvidia cards can reach up to 20k sol/s this is not easy for the CPU. I expect that when properly fixed the above mentioned Celeron may handle 4-6 GPU but if you have a larger rig it would be better mining an other coin on the other GPUs or have a more potent CPU in the rig.

[ALL OS] At slow / unstable Internet connection sometimes lolMiner still submits shares but server does not accept them (Found by me, Confirmed by an offline friend of mine)
Reason: Well - weak connection is always bad for mining. Often re-connect helps.
Problem for Miners: lolMiner does not detect this yet and thus sometimes computes but shares are not counted. So you can not leave the miner unattended.
Expected Time to Fix: Today. I will add a timer that measures the time difference between last submitted share and the last accepted one. Will cause an automatic re-connect if this time is too large.

[Windows / Official AMD drivers] Miner starts, but reports 0 sol/s or Miner does not start up with CL-compile error (Confirmed by me)
Ps: I am confused... I could be certain that someone yesterday gave a Screenshot of this and reported it but now the post is gone. Why?
Reason: Unknown yet.
Problem for Miners: Can not mine.
Expected Time to Fix: I can reproduce the error (when using CPU part on my Kaveri machine), but do not know why this happens yet. Will require deeper investigation because it seems that the OpenCL code does not return what it should do (and what it does on other platforms...)
 
[Windows, Linux (?) / Official AMD drivers] Miner does not start up with CL-compile error (Found by me, Confirmed by qwep1 and ruckus_pro)
Reason: Some problem with the OpenCL code for some AMD OCL drivers. It is unknown yet if this is also true for Linux version. Linux version with Open Source drivers is stable (but slower)
Problem for Miners: Can not mine.
Expected Time to Fix: I can reproduce the problem on some ages old machine at home. I will have a look into the issue later today and hope to get rid of this within next week.

[Windows] Random crash during execution (Reported by AAile, Confirmed by Aeon_)
Reason I do not know yet. It may be that the Nvidia driver things that the App over-utilizes the card. Currently lolMiner has only a very few things done on CPU before the next iteration starts (because the SHA stuff is done in an other thread), so the pressure on the GPU is very high.
Problem for Miners: The miner can not leave the miner unattended.
Expected Time to Fix: I need more info about when this happens and why. I hope that I can get rid out of this with the CPU load fixes above. Else one countermeasure could be running an external demon that checks if the miner is still running and restart it when it does not. Also I consider inserting a parameter that lowers GPU utilization (but of cause this will cost hash-rate... maybe one finds a stable trade of). I thing at least 2 of this countermeasures will be implemented within next two weeks, hopefully earlier. But for a definite fix I have to find the exact reason first


Ps:
I like transparency and so I report that from current dev-fee income lolMiner-mnx has a market share of 0.8 % (Approx 400 ksols from 49 MSol) and I collected 0.055 mnx as def-fee till now. Thats not that much but more will surely come when the miner gets more stable.  I also wanted to announce that I want to pay back approx half of the income of lolMiner-mnx from its first month of existence to those helping me (be detailed reporting) here getting it stable or test it on different platforms during the first two weeks of existence. Of cause I can not say how much that will be in detail yet, but hey, that offer is better than none

Nice miner: On windows 10:

- 23000 s/s on a titan xp (gpu +150, mem +600, power 90%). The miner crash from time to time (I have a loop for automatic restart)

- a little under 15000s/s on a vega 64 (1404 gpu, mem @ 1100, bloackchain driver): but  the card doesn't feel really stressed.

The nvidia thread needs a good dedicated physical core to obtain this hashrate. Hashrate on titan xp  depends on memory speed (25000 possible with mem @1000 and fan very high,  but not stable: videocard freeze the pc with such an overclock on mem).

Nvidia performance is very good (in comparison with optiminer). Amd one is a little underwhelming in comparison with nvidia, considering opencl miner are usually better on amd. There are probably room for optimisation on vega card.

Hi,
Don't know what's wrong but I have the same error:

http://s2.1pic.org/files/2018/02/13/b92153185f39761a06fb.jpg

2 HD7970 2 RX560 drivers 17.10

Really impressive job! Very nice.
I've tried beta for win. On two system crashed but it's a beta.
Rx480 I've got 6k s/s but I've not tried with original bios. Let us know news.
Thank you

I left all niche 6 instances of the miner, in 2 separate rigs. 2 instances in every rig crashed

We cannot use your mine if we cannot leave it unatended

First of all, thank you for your hard work.
Secondly, running even 2 instances on a 7 card nvidia rig will reduce my hash rate. (6x 1070 and 1x 1070ti) It only got as far as 23ksols when I ran 7 instances, they were however constanly crashing as expected. At the moment I am better off running 1 lolminer with 1 card and use ccminer alongside for the other 6 cards. Are you going to be able to patch this up and enable us to run more than one card without the hashpower loss?

Yes indeed. We could go deep into the technical discussion here why, but I try to make the story short. Each shader processor has some shared resources shared within the local work group size threads that you want to adjust. Using more helps the GPU to hide memory latency. That is particularly true for Nvidia Pascal Mainstream Chips with 128 cores in one shader processor. (And here we have a huge difference to Tesla P100 / V100 or AMD Vega GPUs).

In lolMiner we handle things differently by ensuring that shared resources are enough to run with multiple work groups simultaneously. Overall speed is kind of product on how well the chips scheduler can run multiple tasks on one compute core and the way the kernel itself handles memory access. Lets say it this way: I took care about this issue, thats why lolMiner is so fast (for those users without stability issues ^^)

Hopefully tomorrow this time I will have a solution for the build program error you had *crossing fingers*.



Thanks for that numbers  (By the way: which OS?)

For adding CPU mining you must install the OpenCL drivers for your CPU. For the Intel CPUs there is an Intel CPU OpenCL driver at Intel website I used. The drivers for my Kaveri AMD APU came with the installation of the AMD graphics driver, which also added OpenCL support for the CPU.

But honestly: If not really necessary I would not recommend installing the AMD graphic drivers beside the Nvidia ones. That may be more painful then worth the effort, because that Ryzen is very quick, but on CPU we are mainly bounded by memory bandwidth and memory response time. So I won't expect more then about 600 sol/s. Furthermore when your CPU is mining simultaneously to the GPU that may cost GPU hashrate since the calculation against the target, the OpenCL framework and the stratum connection to the pool run on the CPU and I fear that penalty would cost more then the extra hash rate from CPU cores  

Nothing, just added as --device 1. In OC, at the moment, Core - 122, Mem - 350, PL - 85. MSI gtx 1050 ti

What command line do you use? My 1050tu gives 3300 sols/s

Hello, I really do not understand what command to add CPU in miner? got ryzen 7 1700, want to try it.
btw GPU worked good, thank you. by Windows 10
1080 ti 20600 sol/s PL - 80, MEM - 450
1050 ti 6600 sol/s PL - 85, MEM - 350

UPDATE after 12 hours, on WIN 10:
A few problems, 1) got 2 crash 2) An freeze when scrolling bat info to locking some problems, type "A" and freezing wit that. Check pic.

2 - https://i.gyazo.com/c182f2c8a7c2195faeb78c0b8c1c038c.png

With other opencl miner, i noticed that local size of 384/768 speeded up a lot things with nvidia

I have 1060, 1070 and 1070ti with latest nvidia drivers