Topic: PhoenixMiner 6.2c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux) - page 397. (Read 784871 times)

One of my rig which has got mixed cards rx570, 580 and nvidia. version 3 is crashing just before dag creating starts.. i couldn't figure out why? 2.9 works smooth. strange problem

The PhoenixMiner 3.0c always crash at start, windows 10 anniversary, only way to work is to set compatibility to windows 7 on properties, the other versions before this one did not have this problem.

Why it still shows "eth" when mining "etc". I have set the  parameter -coin etc but everything is still showing eth. Otherwise it's running fine on the pool.

Thank you).
and can there be any recommended parameters for cards nVidia that affect stability and efficiency?

That depends of the share difficulty, luck...

My average (reported) speed is +/- 92MH/s. At the pool (nanopool too), it can vary from <50 to >200+MH/s. Let the miner work for 12H minimum, you'll see your effective speed much closer to your average.

Hi
who then can know why the average speed and the effective speed are so different? I have an average speed of 160, and an effective 144 ..
5 * GTX1070
bat - PhoenixMiner.exe -pool etc-eu2.nanopool.org:19999 -wal  -nvidia
pause

I like the miner but it crash on my every time after the dev fee

2018.06.19:20:50:52.316: eths Eth: Received: {"id":6,"jsonrpc":"2.0","result":true}
2018.06.19:20:50:56.320: main Eth speed: 50.146 MH/s, shares: 9/0/0, time: 0:19
2018.06.19:20:50:59.332: main DevFee: Disconnected and stopped
2018.06.19:20:50:59.847: GPU1 GPU1: Starting up... (0)

Any hints?

  Usually it is good idea to put them in the .bat file (or in the config.txt file, if you are using it). In such case you will need to run auto-tune again if you change the clocks on any of the cards, if you install new version of the driver, or if you try using the alternative kernels (-clkernel 2), as the best -gt values could be very different for the two types of kernels. It is also good idea to check if the -gt values are still the best if you upgrade the GPU drivers or install a new version of PhoenixMiner in the future.

   However if you are restarting your rigs only once a week or so, you can just let auto-tune run at each startup as it takes only a few minutes.


Dear PM I was just wondering what is differnece between Kernels 1(optimized)? and 2(alternative)? Are they specific to any type of GPUs?
Thanks

   Depending on what secondary algorithm we do because we once (almost) implemented DCR and it went to non-profitable. It wasn't entirely lost effort but close to it. After finishing the Linux version we will reevaluate the most profitable choices for a secondary coin and then it would take additional month or so. So at least two months in the future.

  Yes. We check if the Compute mode is required by both for each GPU, and the driver. For example, if you are using the blockchain beta driver, the Compute mode is not needed because the huge page mode is already turned on in the driver.

  Usually it is good idea to put them in the .bat file (or in the config.txt file, if you are using it). In such case you will need to run auto-tune again if you change the clocks on any of the cards, if you install new version of the driver, or if you try using the alternative kernels (-clkernel 2), as the best -gt values could be very different for the two types of kernels. It is also good idea to check if the -gt values are still the best if you upgrade the GPU drivers or install a new version of PhoenixMiner in the future.

   However if you are restarting your rigs only once a week or so, you can just let auto-tune run at each startup as it takes only a few minutes.

   If you are getting the infamous 999 error, you need to lower the memory overclock by 10-20 MHz.



According to nanopool documentation, you have to specify the pool like this:

To specify that only the first 9 cards should be used by given PhoenixMiner instance, use -gpus 123456789. To specify the first 11 cards, use -gpus 1,2,3,4,5,6,7,8,9,10,11. The only important thing is to use commas (without spaces) to separate indexes if you are going to use 10 or more as an index of a GPU.


   The clocks are reset when the miner is closing to prevent crashes if you start something else that loads the GPU in different way and the voltage settings are not good enough. Ethash is notoriously tolerant to low core voltage but if you try to run any intensive game or more core-intensive mining algorithm, the GPUs will most probably crash if the settings are left unchanged. It is not that hard to add such option though, and we will add it in the next version, time permitting.

   As for the lower hashrate, please try running both the normal and alternative (-clkernel 2) kernels and run auto-tune with both to see the maximum obtainable hashrate.

   Hopefully in the next release.


   Sorry about this issue. To help us reproduce it and fix it, can you please tell us what -gser option you are using and with what cards/drivers/OS you are using?

   We will add option to avoid resetting OC on exit in the next version.

    You can use the -worker option too depending on the pool.

Hi PhoenixMiner!
There was a problem with the miner on the https://etherdig.net pool with 1080ti graphics cards on port 4444.

 
Miner every 1 hour 23 (25) minutes and not earlier sends errors "Rejected", etc in a circle. If you turn off the miner, then again in 1 hour 23 (25) minutes will go the error "rejected."
This is something related to devfee. This is exactly with the 1080ti cards.
Check, please.

]

Was having same issue right after the Apr 2018 update.  Current rig is 8 1060s with 1000W PSU.   Change swap file to 22GB (22000 MB) and 3.0c started working as expected. 

Thanks for the tip...

oops...  PM died after about 10 min..  Up'd the swap to 25 GB.

try increasing your swap memory to at least 24 gig

I have also 1070 gpus and i using 2x 850W PSU.

My MSI OC Settings are:
Power 65
Temp 76
Core Clock 0
Memory Clock 710 -> The only exception is one gpu because I had problems with this CUDA error. This gpu have memory clock 650 other settings are same.

https://imgur.com/NravSg9.png

Hey Im hoping someone can help me out with an "out of memory" error I keep getting. I have a rig with the following build:
8 X NVIDIA 1070s
Asus H270-P motherboard
2X EVGA 750 PSUs
Windows 10
120 GB harddrive
8GB RAM

I am mining with PoenixMiner and using Nanopool and keep getting an error message.  (I have 2 other rigs built the exact same and havent had any issues with them).  Yesterday it was mining fine and the the program launched out, so I reset the rig and have been having issues since.  I was able to get 7 GPUs running before but when I plugged in the 8th card, its giving me an error code.  I tried switching out the risers and now am only able to get 6 cards working before the same error occurs  -- "CUDA error in CudaProgram.cu:377 : out of memory (2)"

MSI Settings:
Power: 64
Temp: 67
Core Clock: 0
Memory Clock: 585

Any ideas??

My experience with Phoenixminer 3.0.c is that it gives a lot of malfunctions than 2.9 version, miner starts spontaneously due to malfunctions, so I lower afterburner settings, obviously it costs more energy in this way.

It is safe, you can ignore the warning.

An honest advice though, better read and research more before starting to mine.

It is clear from your question that you can be an easy hack target or even destroy your hardware yourself.

Cheers

Win10 Pro detect the software with trojon virus, pls advise

StreaMex,

You need to add a return or newline at the end of your code.

here is an example of my code:

We're trying to integrate this new miner with our python monitoring system, and it does not seem to receive any socket response.

Documentation for PhoenixMiner claims that it uses same api approach as Claymore.

Here's our code:


It works perfectly well with Claymore, but it gets stuck at "Socket receiving response..." with Phoenix miner. So it seems it does not receive any data from socket connection. I could not find any detailed documentation on Phoenix's api, other than a statement that "it is the same as Claymore protocol".

Please, help - did I miss anything in documentation, is there a more complete documentation or how should I fix the code to work with Phoenix?

Thanks for your time and attention.

I have 2 ETH rigs, one with 6x1070s, one with 6xRx570s.  All my 1070s are Nvidia branded FE and set to power target = 65% with NvidiaInspector. All hash at 32.5-33 Mh/s and consume 99-100 watts.  My Rx570s are a mix of Asus and MSI cards, all are bios modded.  The Asus cards consume 92-94 watts, MSI cards range from 95-105 watts and all 6 hash at 29.0 Mh/s.  So for me Nvidia cards are more efficient hash/watt, but cost a lot more (now) than AMD cards.  Nvidia cards are very easy to tune, AMD cards CAN be easy (find a good bios online and flash it) or a pain (roll your own bios mod and spend hours tweaking).  For my ASUS cards, I found a great BIOS on Anorak, flashed it and I was off and running.  My MSI cards didn't like any of the straps I tried, so spent a lot of time trying to find an optimal setting.  Also have 3 of the exact same MSI cards with the exact same bios, same memory, etc, and they use 95w, 99w, 105w.