Genesis Mining Presents: SGMiner-GM - now with Zawawa's GG! [Updated 17/01/2017]

[moonmaths]

Which are the usual scores with RX470/80 on monero

my experience (xmr only here)
rx480 - 4gb Sapphire nitro+ oc:
Samsung Memory non-flashed: 650 h/s
Elpida/Hynx Memory non-flashed: 580 h/s
with rom flashed (just the simple copy/paste of straps and overclocking to around 2040, nothing fancy)
Samsung: 730-750 range
Elpida/hynx memory: up to 710 to 730 range

Note: I'm not sure exactly what the other memory type is, but the 'other' polaris bios editor works with it (check in comments of the github page). 
Also, using Ubuntu 16.04, not windows..

[1715154640]

1715154640

[1715154640]

1715154640

[1715154640]

1715154640

[1715154640]

1715154640

[manotroll]

I would like to know which values to use to mine etc with the RX 480?

How to get the best income for Groestlcoin?

[sammy007]

Is there any stratum spec available for this ETH miner?

[avirdoz]

Maybe someone can point me - what mean l in binary file name:
cryptonightTongagw256l4.bin
w256 - worksize 256
but l4?

[citronick]

Maybe someone can point me - what mean l in binary file name:
cryptonightTongagw256l4.bin
w256 - worksize 256
but l4?

Intensity?

[Slydrule]

Maybe someone can point me - what mean l in binary file name:
cryptonightTongagw256l4.bin
w256 - worksize 256
but l4?

I could be wrong but I think it designates the OS.  Windows bins will have l4.  Linux l8.  I would not bet the ranch on this, but it is an observation of mine.

[avirdoz]

Maybe someone can point me - what mean l in binary file name:
cryptonightTongagw256l4.bin
w256 - worksize 256
but l4?

Intensity?

It's lowercase L, so i don't think so.

[Elun]

It is just size of long on your platform.
https://github.com/genesismining/sgminer-gm/blob/master/ocl.c#L343-L347

[mankeen]

Is this one faster than Claymores miner?

[yoyo1]

Hello.
I wonder why not to use Wolf kernel in this sgminer?
As far as I can see there is a small thread related to the case when unmodified cards achieve great hashrate with very moderate consumption:
http://forum.ethereum.org/discussion/comment/45601#Comment_45601

[wacko]

Is it possible to build this miner for Mac OS? Thinking about running it in esxi virtualized with some older GPU like 7870/7950/7970 passed through, would be nice to mine XMR on that system as well. Or is it only for Windows and Linux?

[avirdoz]

Is it possible to build this miner for Mac OS? Thinking about running it in esxi virtualized with some older GPU like 7870/7950/7970 passed through, would be nice to mine XMR on that system as well. Or is it only for Windows and Linux?

https://github.com/genesismining/sgminer-gm/blob/master/doc/FAQ.md#can-i-mine-with-sgminer-on-a-mac

[wacko]

Is it possible to build this miner for Mac OS? Thinking about running it in esxi virtualized with some older GPU like 7870/7950/7970 passed through, would be nice to mine XMR on that system as well. Or is it only for Windows and Linux?

https://github.com/genesismining/sgminer-gm/blob/master/doc/FAQ.md#can-i-mine-with-sgminer-on-a-mac

Thanks! Not sure how I missed it, was googling for a bit about mining on Mac OS before asking here and didn't find that page for some reason.

sgminer will compile on OSX, but the performance of GPU mining is compromised due to the OpenCL implementation on OSX, there is no temperature or fanspeed monitoring, and the cooling design will usually not cope with constant usage leading to a high risk of thermal damage. It is highly recommended not to mine on a Mac.

Hmm, no temp and fanspeed monitoring is inconvenient but I can live with that. "Cooling design" and "thermal damage" stuff is not relevant I suppose since it's going to be a regular undervolted Tahiti or Pitcairn graphics card in my home server's full tower case.. But the "compromised performance due to OpenCL implementation on OSX" doesn't sound good. I wonder how much compromised it is? Is it like 10% lower hashrate, or 50% less?

[scryptr]

Does anyone know how to make sgminer-GM work on nicehash?

It gets 100% rejected shares.

On any other pool it works perfect. (eg nano or dwarf)

There is a sgminer-gm v5.5.5-4 by NH team, that is suppose to fix the cryptonight issues..... but I still get 100% rejects too (under Linux). Any ideas?

NICEHASH HAS RELEASED UPDATES--

The latest update is SGminer-GM-5.5.5-8, modified to work on the NiceHash pool.  My finicky RX 460 cards run smoothly on it, as do my RX 470 cards.  Here is my configuration file:

<=======CUT LINE=======>

{
    "pools": [{
        "url": "cryptonight.usa.nicehash.com:3355",
        "user": "YOURBTCADDRESS.WORKER",
        "pass": "x"
    },
    {
        "url": "cryptonight.eu.nicehash.com:3355",
        "user": "YOURBTCADDRESS.WORKER",
        "pass": "x"
    }],
    "profiles": [{
        "worksize": "8",
        "name": "xmr",
        "algorithm": "cryptonight",
        "gpu-threads": "1",
        "rawintensity": "1016"
    }],
    "default-profile": "xmr",
    "no-extranonce": false,
    "no-submit-stale": true,
    "api-listen": true,
    "api-port": 4028,
    "api-mcast-port": 4028,
    "api-allow": "W:127.0.0.1,W:192.168.1.1/24",
    "temp-cutoff": "85",
    "temp-overheat": "80"
}

<=======CUT LINE=======>
 
Don't include the "CUT LINE" notes, of course.  The "api-port" lines allow for monitoring software.  The raw intensity is less than 1024 per the software authors' (Wolf0's crew) specifications.  The "no-extranonce" line is set to "false" because NiceHash uses extranonce.  The "worksize" of "8" is suitable for one GPU thread, or a "worksize" of "4" can be used for two GPU threads.  The worksize is very small compared to many other algorithms mineable by SGminer softwares.  Adjust values to your preferences.  Call the configuration with "sgminer -c start.conf" from within a BASH script or Windows batch file.  I am sure you know how to set the environment variables either way.

This miner runs for weeks on my most tempermental cards.  Hope this note helps.       --scryptr

EDIT: Please note that this edit is in September 2017!

After making some slight changes to the above configuration, it represents my current configuration.  I have increased the Raw Intensity from 1008 to 1016, and included the term "no-submit-stale" as true.  This squeezes about 1-2 more hashes per card, and removes almost all rejects.

CryptoNight is currently a top earner on WhatToMine, and SGminer-GM is mining well on my rigs.  Some rigs have gone several days with no rejects or hardware errors.  Thanks again to the authors and sponsors of this software!       --scryptr

[citronick]

Does anyone know how to make sgminer-GM work on nicehash?

It gets 100% rejected shares.

On any other pool it works perfect. (eg nano or dwarf)

There is a sgminer-gm v5.5.5-4 by NH team, that is suppose to fix the cryptonight issues..... but I still get 100% rejects too (under Linux). Any ideas?

NICEHASH HAS RELEASED UPDATES--

The latest update is SGminer-GM-5.5.5-8, modified to work on the NiceHash pool.  My finicky RX 460 cards run smoothly on it, as do my RX 470 cards.  Here is my configuration file:

<=======CUT LINE=======>

{
    "pools": [{
        "url": "cryptonight.usa.nicehash.com:3355",
        "user": "YOURBTCADDRESS.WORKER",
        "pass": "x"
    },
    {
        "url": "cryptonight.eu.nicehash.com:3355",
        "user": "YOURBTCADDRESS.WORKER",
        "pass": "x"
    }],
    "profiles": [{
        "worksize": "8",
        "name": "xmr",
        "algorithm": "cryptonight",
        "gpu-threads": "1",
        "rawintensity": "1008"
    }],
    "no-extranonce": "false",
    "default-profile": "xmr",
    "api-listen": true,
    "api-port": 4028,
    "api-mcast-port": 4028,
    "api-allow": "W:127.0.0.1,W:192.168.1.1/24",
    "temp-cutoff": "85",
    "temp-overheat": "80"
}

<=======CUT LINE=======>

Don't include the "CUT LINE" notes, of course.  The "api-port" lines allow for monitoring software.  The raw intensity is less than 1024 per the software authors' (Wolf0's crew) specifications.  The "no-extranonce" line is set to "false" because NiceHash uses extranonce.  The "worksize" of "8" is suitable for one GPU thread, or a "worksize" of "4" can be used for two GPU threads.  The worksize is very small compared to many other algorithms mineable by SGminer softwares.  Adjust values to your preferences.  Call the configuration with "sgminer -c start.conf" from within a BASH script or Windows batch file.  I am sure you know how to set the environment variables either way.

This miner runs for weeks on my most tempermental cards.  Hope this note helps.       --scryptr

thanks scryptr.

SGminer-GM-5.5.5-7 was the first version that worked on CN algo in NH, before that all rejects.

I will try out SGminer-GM-5.5.5-8 soon -- hope good performance too.

Too bad Cryptonight profitability in NH took a nose dive of late... 

[galaxy5user]

Hi, everybody! After I have been disappointed in Claymore's GPU miner (who was too lazy to backup GNU/Linux sources and too greed to buy a secure online storage, like, for example, lessauthority). I started too look for an open source GNU/Linux-compatible alternative. SGMiner looks decent enough. I was able to run the v5.5.5 binary for Ubuntu (and not able to build from sources yet).

What I don't understand is how to view GPU's temperature and FAN speed? When I run "sgminer_ubuntu64 --ndevs" it shows me:
"Unable to load ATI ADL library". I found the library is being loaded by the function:

Code:

dlopen( "libatiadlxx.so", RTLD_LAZY|RTLD_GLOBAL)

AMD ADL API docs says the file must reside in the following directory:

Linux 64-bit Runtimes (64-bit OS):
/usr/X11R6/lib64/Libatiadlxx.so

But the file is not found anywhere on the system!
I have the driver "amdgpu-pro-16.60" from the official site under Ubuntu 16.04.

Also, where I can find the meaning of the environment variables which are being complained to?:

WARNING: GPU_MAX_ALLOC_PERCENT is not specified!
WARNING: GPU_USE_SYNC_OBJECTS is not specified!

[Arnoldent]

Hi, everybody! After I have been disappointed in Claymore's GPU miner (who was too lazy to backup GNU/Linux sources and too greed to buy a secure online storage, like, for example, lessauthority). I started too look for an open source GNU/Linux-compatible alternative. SGMiner looks decent enough. I was able to run the v5.5.5 binary for Ubuntu (and not able to build from sources yet).

What I don't understand is how to view GPU's temperature and FAN speed? When I run "sgminer_ubuntu64 --ndevs" it shows me:
"Unable to load ATI ADL library". I found the library is being loaded by the function:

Code:

dlopen( "libatiadlxx.so", RTLD_LAZY|RTLD_GLOBAL)

AMD ADL API docs says the file must reside in the following directory:

Linux 64-bit Runtimes (64-bit OS):
/usr/X11R6/lib64/Libatiadlxx.so

But the file is not found anywhere on the system!
I have the driver "amdgpu-pro-16.60" from the official site under Ubuntu 16.04.

Also, where I can find the meaning of the environment variables which are being complained to?:

WARNING: GPU_MAX_ALLOC_PERCENT is not specified!
WARNING: GPU_USE_SYNC_OBJECTS is not specified!

I do not know about Linux. But in the Windows, it is better to set:
setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 100

[galaxy5user]

I obtained that there is possibility to set the fan's speed manually by writing to the following file:
echo 220 > /sys/class/drm/card0/device/hwmon/hwmon0/pwm1
For me the command returns "Permission denied" even though I'm calling it with sudo. Any suggestions?
UPD: I have workarounded it by using 'tee' command.

[galaxy5user]

Is there any comprehensive guide for selecting 'rawintensity' and 'worksize' params? What I have found is people trying to 'hit the sky' and exhange each other with 'good' (proven) values.

[Elun]

Is there any comprehensive guide for selecting 'rawintensity' and 'worksize' params? What I have found is people trying to 'hit the sky' and exhange each other with 'good' (proven) values.

worksize - it is ocl localworksize of your workgroup. 64 is kind of optimal for amd on most algos(not for equihash i think, this algo is more complex than just compute hash). For amd it will vary from 1 to 256. Best sizes if it is divisible by 64. Or at least 2. But again it depends from algo and its implementation.
rawintensity - when miner compute hashes, it check nonce from 0x00000000 to 0xFFFFFFFF. You can't say to gpu: compute this whole range at once, that wouldn't be effective. So you should split this range in pieces(in this case it would be global worksize in opencl terms. You can think about it like a number of hashes that sgminer would compute before verifying gpu's work on cpu and moving on to a next piece of nonces). rawintensity would be the size of such piece. It can be selected based on hardware properties of your videocard, or to be chosen experimentally(this is what most miners do, lol). xIntensity or intensity arguments allow you to set this rawintensity without computing it by yourself, as i remember intensity will set globalworksize to 2^intensity. I don't know if you can apply this to equihash impl in sgminer, i'm talking based on exp with other algos. I think Zawawa can tell more, or someone who ported his impl into sgminer-gm.

UPD: actually, here is your guide: https://github.com/genesismining/sgminer-gm/blob/master/doc/MINING.md
with intensity and worksize description.