MultiPoolMiner / MultiPoolMiner

Monitors crypto mining pools in real-time in order to find the most profitable for your machine. Controls any miner that is available via command line.
GNU General Public License v3.0
898 stars 343 forks source link

MPM failed to connect to miner (PalginNvidia) #991

Closed FruitFace closed 6 years ago

FruitFace commented 6 years ago

The new PalginNvidia miner does run, and accepts shares, etc, but MPM gives the error in the title. Eventually, the watchdog fails it out.

2018-01-09_21-16-03.txt

It does complain about a wrapper file, not existing in the wrapper subfolder. I did copy the whole directory to a folder off the root of C, (path with no spaces) to try and see if it was related to the wrapper path bug, but it still complained.

UselessGuru commented 6 years ago

For this miner you will have to use one reliably you need to disable the watchdog.

Reason: PalginNvidia does not have API so MPM must attempt to read the hash rate from the screen output (that's what wrapper is for). However PalginNvidia sometimes takes ages to report hash rates. Watchdog kicks it out of the list of available miners.

BTW: Palgin has announced that API implementation is in the works :-)

cmbconcretefire commented 6 years ago

Just what I was looking for cause from my research tonight, I believe this Palgin-Nvidia miner for neoscrypt blows the other neoscrypt miners out of the water. I did some testing and it's fast.

cmbconcretefire commented 6 years ago

Hey @UselessGuru , I disabled watchdog, restarted MPM, and I'm still getting the failed to connect to PalginNvidia. Am I missing something?

Braintelligence commented 6 years ago

@UselessGuru And I thought my swap file size was too small <_>. Good to know it's not my fault then.

UselessGuru commented 6 years ago

I disabled watchdog, restarted MPM,

Disabling WD will prevent that pools or miners are removed from the list. WD has no impact on the timing loops built into the API files (wrapper included). You will still see the errors. Live with it :-(

cmbconcretefire commented 6 years ago

@UselessGuru , where are you from? I'm in the US. Just curious. So, are you saying I will still see the errors but that it will actually report to the pools correctly? To be honest, I can't even get it to finish 'benchmarking'.....or get it to report to zpool.

UselessGuru commented 6 years ago

where are you from?

Switzerland

it will actually report to the pools correctly

yes, it sends hashes to the pool.

or get it to report to zpool?

??? Do you see any errors in the miner windows? If not then it will send hashes to the pool. Note that it takes a few minutes of mining until you will see the statistics getting updated at the pool.

I can't even get it to finish 'benchmarking'....

As written before Palgin takes ages to report a usable hash rate on screen that the wrapper can read.

titvs commented 6 years ago

The API implementation and the support for GTX 9xx would be great. I have a machine with a GTX 980 that cant be wasted... ;P

FruitFace commented 6 years ago

@UselessGuru Thanks for the detail. Disabling the watchdog did the trick. The PalginNvidia is the best performer for me, now, that the original Palgin miner is no longer available. I get about 45-57MH more than CCminerKlaust on Neoscrypt, the next best Neoscrypt miner, at least for me.

I'm mining to NH, running 3x 1070s and 1x 1080, and Neoscrypt is generally my most profitable algo, followed by Lyra2rev2, Equihash and recently, Nist5.

The miner is stable as well, as Klaust would sometimes crash one of my 1070's, if i was using firefox, despite the 1070 being one of the three using a PCI extender. The 1080ti is my main vid card, as this is a multi-use system Go figure...

BTW, I am using switching prevention 0, since NiceHash is PPS. It does switch way more, now, though I can't really tell, yet if it has helped, or hurt. I guess I'll need several more cycles, before I can see if there's a trend. But so far, the change appears to have been fairly neutral to earnings.

torsrex commented 6 years ago

I noticed that in the latest release of PalginNvidia they changed the way the hashrate is printed to this INFO : [20:00:40] : GPU#0 - GeForce GTX 1050 Ti, speed is 520.94kH/s Line 35 in wrapper.ps1 looks like this If (($Line -like "*total*" -or $Line -like "*accepted*" -or $Line -like ">*") -and $Line -like "*/s*") { But none of the words matches the new printing layout of the miner. If however the line is changed to this if (($Line -like "*total*" -or $Line -like "*accepted*" -or $line -like "*speed*" -or $Line -like ">*") -and $Line -like "*/s*") { It successfully reads the output of the miner. Perhaps this could be updated?

PapaChealio commented 6 years ago

That particular output text is done per card. MPM is interested in the aggregate hash rate from all the cards, which this miner only currently outputs every 20 accepted shares or so. The "wrapper" picks this up aggregate output from the "Share accepted, " text. Unfortunately, this output doesn't happen often enough to avoid getting flagged by the watchdog once the share submission rate slows down (i.e. pool changes to harder stratum difficulty).

If you've only got one card, you could look for that one line but PalginPav is supposedly going to fix this issue in 1.0.2 anyway.

grantemsley commented 6 years ago

Closing outdated issue. If this is still a problem (and there currently isn't an open issue for it) please open a new one.