XMR-stak-JK 2.10.7 Compiled with no devfee (Updated: 2019-8-18)

[JaredKaragen]

Followed all instructions on the XMR-stak github to compile.... kinda.  Notes are below.

Huge thanks to Fireice-UK for making this source public.  Without it;  this release would not be here.  If you like it;  I suggest you donate to him directly rather than the devfee as it costs you more than just the donation time of hashrate in the real scope of things.


Compiled using Cuda 10.  
DevFee set to zero.  
Build version number stayed the same, build name was changed (added -JK).  
No major modifications aside from "built by notes", and changing the donate address to '127.0.0.1' for quick timeouts if there is any code trickery I did not catch.  CLi.cpp file was modified to add built by notes, as well as remove excessive fluff.

No donations or whatever required.  Just enjoy.  I know a bunch of you (like me) would like the zero devfee build, but either can't or are too lazy to compile it yourself.   Here you go you lazy ones  If you do feel like taking the time to donate to me for my time, you can find the donation address in the "miner batch" post that is linked in my signature just message me...  But I highly suggest throwing a donation to fireice-uk first... without him, this release would not be possible.

***I literally cut/paste the source and built it 100% as per instruction aside from the edits mentioned above.***

MS Visual C++ Runtime:
Download here

Current Version:
Download XMR-Stak-2.10.7-JK (VS2019 Build) here
Download XMR-Stak-2.10.7-JK here

Previous versions:
Download XMR-Stak-2.10.6-JK here
Download XMR-Stak-2.10.5-JK here (NOTE: suggested to delete old .txt files and let them get rebuilt due to some changes)
Download XMR-Stak-2.10.3-JK here
Download XMR-Stak-2.10.2-JK here
Download XMR-Stak-2.10.1-JK here
Download XMR-Stak-2.10.0-JK here
Download XMR-Stak-2.9.0-JK here
Download XMR-Stak-2.8.3-JK here
Download XMR-Stak-2.8.2-JK here
Download XMR-Stak-2.7.1-JK here
Download XMR-Stak-2.6.0-JK here
Download XMR-Stak-2.5.2-JK here
Download XMR-Stak-2.5.1-JK here
Download XMR-Stak-2.4.4-JK here
Download XMR-Stak-2.4.2-JK here

If you are getting slower hashrates on older cards, try these DLL files.  They are from a different release of xmr-stak, thus not compiled by me.  But, another user has reported that replacing the DLL's with these has increased performance.  Your mileage may vary:  https://mega.nz/#!oGY0iIxa!X1tlll_HR5LWTa1aT3jBQIZAR3Bv91_Bgll8huVl84w  This was applicable on 2.4.x only.

*update*
2019-08-18: Compiled and uploaded 2.10.7 under VS2019
2019-08-12: Compiled and uploaded 2.10.7
2019-07-23: Compiled and uploaded 2.10.6
2019-06-15: Compiled and uploaded 2.10.5
2019-03-28: Compiled and uploaded 2.10.3
2019-03-16: Compiled and uploaded 2.10.2
2019-03-12: Compiled and uploaded 2.10.1
2019-03-07: Compiled tested and uploaded 2.10.0
2019-03-04: Compiled and uploaded 2.9.0
2019-02-16: Compiled and uploaded 2.8.3
2019-02-15: Compiled and uploaded 2.8.2
2018-12-10: Compiled and uploaded 2.7.1
2018-11-23: Compiled and uploaded 2.6.0
2018-10-29: Compiled and uploaded 2.5.2
2018-10-18: Compiled tested and uploaded 2.5.1
2018-06-03: Compiled and uploaded 2.4.4
2018-05-30: Added vc redistributable install link
2018-05-29: Compiled and uploaded 2.4.3/quickly removed partially built binary
2018-04-10: Alternate .dll files from original dev's older release available, download link added.  May give better or worse hashrates.
2018-04-06: Fixed issue with not launching properly.  Added original usage docs from source.

Just run xmr-stak.exe and fill out the questions.  from there after, just run xmr-stak.exe and theres no need to put any commands after.
See my below replies for notes on disabling CPU and GPU's individually.  I can help if you need assistance with your configuration.

Please see all of the build and usage notes on the original github: By clicking here.



If you are looking for XTL-stak, that's right here.

[1714016878]

1714016878

[1714016878]

1714016878

[tbearhere]

Hi JaredKaragen....how you been?
Thx for your mining app.

EDIT: There is no read me file....I'll try using

:start
rem for 750ti -l 8x30
set xmrAddress=addy
set poolPassword=x
set poolUrl=stratum+tcp://poolxxx:1234
set launchParam=-l 8x30
rem greater value means more interactivity of your system but lower hash rate
set bFactor=10
rem time between a kernel start
set bSleep=100
xmrMiner_xmr-stak.exe --url=%poolUrl% -u %xmrAddress% -p %poolPassword% %launchParam% --bfactor=%bFactor% --bsleep=%bSleep%
goto start

Hope this works.

[Dentidan]

Is that program faster than the Claymore miner for the R9 290 or 390?

[JaredKaragen]

Just run xmr-stak.exe and answer questions it asks.  It will configure all devices it finds.  There are flags in the xmr-stak docs that can allow you to disable one or the other type of hardware such as cpu.

I'll have to re-compress the archive with the xmr-stak originals; yes, I did not include them.  I'll try to get on that asap.

I'll post again shortly, as well as update original post's download link.

A big storm is coming through Cali and it's messing with my connection.  Please bear with me.

[JaredKaragen]

Is that program faster than the Claymore miner for the R9 290 or 390?

I am unsure.  presently I only have an Rx560 to test and it sees an average of 350H/sec...  which seems on par with recent releases; I started it up ~8 hours ago, so the hashrate average is a real average.

Hi JaredKaragen....how you been?
Thx for your mining app.

EDIT: There is no read me file....I'll try using

I've... been... busy... to say the least.   Still staying afloat and feeding myself.... tis what counts

xmr-stak is a bit different than ccminer.  I have the original post updated, as well as the new link to the replacement files.  There was corruption in one of the DLL's apparently from my thumbdrive.   I noticed it when installing onto one of my machines that normally doesn't run XMR just a minute ago.   Double checked this version.  Worked like it should....  sorry for the issues.


there's flags you can use at launch if you run xmr-stak --help from a command prompt/powershell.  don't quote me (been up over 24 hours now) but its something to the tune of --disablenvidia --disablecpu, etc....

or if you know code somewhat well; and don't want to look up what the flags are for at launch; you can go into the text files generated in the xmr stak folder; and edit the files you wish to do;  if you want to disable a specific nvidia GPU;  just go in and comment out the details for the GPU like this:

Code:

/*
 * GPU configuration. You should play around with threads and blocks as the fastest settings will vary.
 * index         - GPU index number usually starts from 0.
 * threads       - Number of GPU threads (nothing to do with CPU threads).
 * blocks        - Number of GPU blocks (nothing to do with CPU threads).
 * bfactor       - Enables running the Cryptonight kernel in smaller pieces.
 *                 Increase if you want to reduce GPU lag. Recommended setting on GUI systems - 8
 * bsleep        - Insert a delay of X microseconds between kernel launches.
 *                 Increase if you want to reduce GPU lag. Recommended setting on GUI systems - 100
 * affine_to_cpu - This will affine the thread to a CPU. This can make a GPU miner play along nicer with a CPU miner.
 * sync_mode     - method used to synchronize the device
 *                 documentation: http://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__DEVICE.html#group__CUDART__DEVICE_1g69e73c7dda3fc05306ae7c811a690fac
 *                 0 = cudaDeviceScheduleAuto
 *                 1 = cudaDeviceScheduleSpin - create a high load on one cpu thread per gpu
 *                 2 = cudaDeviceScheduleYield
 *                 3 = cudaDeviceScheduleBlockingSync (default)
 *
 * On the first run the miner will look at your system and suggest a basic configuration that will work,
 * you can try to tweak it from there to get the best performance.
 *
 * A filled out configuration should look like this:
 * "gpu_threads_conf" :
 * [
 *     { "index" : 0, "threads" : 17, "blocks" : 60, "bfactor" : 0, "bsleep" :  0, 
 *       "affine_to_cpu" : false, "sync_mode" : 3,
 *     },
 * ],
 * If you do not wish to mine with your nVidia GPU(s) then use:
 * "gpu_threads_conf" :
 * null,
 */

"gpu_threads_conf" :
[
/*  // gpu: GeForce GTX 1070 architecture: 61
  //      memory: 6789/8192 MiB
  //      smx: 15
  { "index" : 0,
    "threads" : 28, "blocks" : 45,
    "bfactor" : 6, "bsleep" :  25,
    "affine_to_cpu" : false, "sync_mode" : 3,
  },
*/
  // gpu: GeForce GTX 980 architecture: 52
  //      memory: 3370/4096 MiB
  //      smx: 16
  { "index" : 1,
    "threads" : 12, "blocks" : 48,
    "bfactor" : 6, "bsleep" :  25,
    "affine_to_cpu" : false, "sync_mode" : 3,
  },

],

notice the /* at the beginning of one GPU section, and the */ closing it;  so my XR stak would now only mine with my second GPU (the 980).

or another example if you dont want to mine on your CPU (cpu.txt):

This:

Code:

"cpu_threads_conf" :
[
    { "low_power_mode" : false, "no_prefetch" : true, "affine_to_cpu" : 0 },
    { "low_power_mode" : false, "no_prefetch" : true, "affine_to_cpu" : 2 },

],

Becomes this:

Code:

"cpu_threads_conf" :
[
   /* { "low_power_mode" : false, "no_prefetch" : true, "affine_to_cpu" : 0 },
    { "low_power_mode" : false, "no_prefetch" : true, "affine_to_cpu" : 2 },
  */
  //  or you can comment out a line one at a time like this line
],

on my machines, I just put a link to the xmr-stak.exe into the system startup folder....   the text/json config files make it so you dont need to launch the miner with the same details every time like ccminer does.


FYI this works on old GTX4xx cards up to the current pascal cards.  only tested with an RX560 as per above, but im sure its got a lot of backwards compatibility considering how many devices it can handle presently.


If you add more gpus to the system or change the processor, etc... just delete the text/json file pertaining (pool/cpu/amd/nvidia) and it will just re-check or re-ask questions it needs to next time you launch it.  If all of your systems are similar, you can copy/paste the JSON configs, and just go into each miner's pool.txt config and change the worker name.


if you are using nanopool, be sure to follow their instructions for connecting with a worker name;  don't reply on the worker name it asked for you separate,  Im not sure exactly yet wht this version asks you if its already included in the username.

[Documan]

I wonder if it works with RX 550 cards. The claymore miner does not work with that.

[RIGED]

will give it a whirl,
using Cast XMR with 1.5% fee seems stupid when you can keep that fee by using your compiled Stak

we are also doing things check it out, https://vegatoken.org/

thx again

[Iamtutut]

Thanks for the release !

[Sgsg666]

Yay! Thanks this is awesome!
Other cryptonight miners you might want to configure:

https://bitcointalk.org/index.php?topic=3058014.0
https://github.com/xmrig/xmrig
https://github.com/Bendr0id/xmrigCC

[JaredKaragen]

Yay! Thanks this is awesome!
Other cryptonight miners you might want to configure:

https://bitcointalk.org/index.php?topic=3058014.0
https://github.com/xmrig/xmrig
https://github.com/Bendr0id/xmrigCC

is there a reason to recompile more than one XMR app?   I mean;  I just wanted to provide a legitimate outlet to a no devfee miner for my miner batch/packages I am assembling.  I liked XMR-stak because it incorporated Nvidia as well as AMD cards;  as well as supporting all my pre-CUDA7 devices. (470, 550, 550ti, 650, 650ti, etc etc.)

Something I am mainly on the look out for is:  an all-around good version of SGMiner that handles everything from X11->X16R and everything in between for AMD cards.   I don't have my hands on a stable and reliable "AMD 'other algo' miner" as of yet.  Same for equihash and ETH miners.  ETH is bottom of my list as it is its own animal; and not a coin/platform I ever believed in.  I only use ETH on exchanges for selling otherwise unsellable coins at a profit.   If I can not make a profit on DOGE markets;  usually I can squeeze a profit out of going from ALT->ETH->BTC.  Since ETH doesn't scale as small as DOGE, its easier to pinch out a modest profit.

I already have the nvidia side almost handled;  I just want to find a few more no fee sources like for equihash and possibly one for ETH.

I wonder if it works with RX 550 cards. The claymore miner does not work with that.

I've already tested with a rx560, so I don't see why it wouldn't work;  the algo isn't super dependent on memory size AFAIK.



Side note:  With this new no-devfee recompiled version;  it has been maintaining a more steady hashrate report at the pool.  fewer wild swings, etc. as well as a better monthly income as per nanopool's estimate.   I think it is due in part to the miner app disconnecting;  and reconnecting to the pool and getting auto-diff reset when a block change happens during devfee mining.   The first few lower diff shares tend to get a fake hashrate bump etc.   I see it now.   Devfee mining really hurt the miners profits in the end....   Maybe you guys will see it eventually.

I wish people treated mining apps as people treated software in the early 90's on BBS's......  We would be a better society as a whole.

FYI: sorry for the damaged first upload.  I had no clue until something was mentioned.   If anyone had downloaded when I first uploaded;  I suggest replacing it with the current posted version at the top of this thread.

[RIGED]

thnx!!

I couldn't get it to run two threads on the vega64
could you please post the code for 6 gpus if you know

thnx again for the effort

[JaredKaragen]

thnx!!

I couldn't get it to run two threads on the vega64
could you please post the code for 6 gpus if you know

thnx again for the effort

Not sure what you mean.

If you have multiple GPU, you will see a config file like this for nvidia:

Code:

"gpu_threads_conf" :
[
  // gpu: GeForce GTX 1070 architecture: 61
  //      memory: 6789/8192 MiB
  //      smx: 15
  { "index" : 0,
    "threads" : 28, "blocks" : 45,
    "bfactor" : 8, "bsleep" :  100,
    "affine_to_cpu" : false, "sync_mode" : 3,
  },
  // gpu: GeForce GTX 980 architecture: 52
  //      memory: 3370/4096 MiB
  //      smx: 16
  { "index" : 1,
    "threads" : 12, "blocks" : 48,
    "bfactor" : 6, "bsleep" :  25,
    "affine_to_cpu" : false, "sync_mode" : 3,
  },

],

You can see each GPU listed with it's configs.   There are notes that I did not cut and paste about modifying 'bfactor' and 'bsleep' to 8 and 100 respectively.  This has helped me in my issues in the past.  the AMD file should contain a similar description of its settings.  I am thinking this is probably your issue?  You should see all system GPU's listed in the AMD.txt file.

I don't have any vegas on hand to test with or anything, so I can't get a config from a vega to show a prime example as the app creates the configs itself... so give it a shot;  assuming I understood your initial question properly

Also;  if you add more GPU's to the system, or remove some, or change the CPU, delete the appropriate file (amd.txt, cpu.txt or nvidia.txt) and re-run xmr-stak.exe.  it will re-assess the system and rebuild the configs for the current hardware set.

[minahan24]

OP, what settings do I need to tweak to increase my speed. I have 6x rx570 and I've run SRBMiner and getting 5700h/s, while I'm only getting 4000h/s with this miner. Maybe you can point me to the settings I need to play around.

[JaredKaragen]

OP, what settings do I need to tweak to increase my speed. I have 6x rx570 and I've run SRBMiner and getting 5700h/s, while I'm only getting 4000h/s with this miner. Maybe you can point me to the settings I need to play around.

I believe it has something to do with the threads and blocks variables;  the .txt file should have an explanation of each one and what it does.

does SRBMiner apply an overclock to the cards when it launches?  that may be the difference in hashrate.


Im not sure is xmr-stak is the quickest miner app overall TBH... but, I do know so far it has a wide range of compatibility with devices compared to most other apps i've tried.

[minahan24]

OP, what settings do I need to tweak to increase my speed. I have 6x rx570 and I've run SRBMiner and getting 5700h/s, while I'm only getting 4000h/s with this miner. Maybe you can point me to the settings I need to play around.

I believe it has something to do with the threads and blocks variables;  the .txt file should have an explanation of each one and what it does.

does SRBMiner apply an overclock to the cards when it launches?  that may be the difference in hashrate.


Im not sure is xmr-stak is the quickest miner app overall TBH... but, I do know so far it has a wide range of compatibility with devices compared to most other apps i've tried.

The SRBMiner doesnt apply any overclock, I have to manually apply it myself. But I'm using the same overclock I've used on the SRBMiner on this miner but getting lower speed so I think I need to play around with the settings. I will update.

[RIGED]

thnx!!

I couldn't get it to run two threads on the vega64
could you please post the code for 6 gpus if you know

thnx again for the effort

Not sure what you mean.

If you have multiple GPU, you will see a config file like this for nvidia:

You can see each GPU listed with it's configs.   There are notes that I did not cut and paste about modifying 'bfactor' and 'bsleep' to 8 and 100 respectively.  This has helped me in my issues in the past.  the AMD file should contain a similar description of its settings.  I am thinking this is probably your issue?  You should see all system GPU's listed in the AMD.txt file.

I don't have any vegas on hand to test with or anything, so I can't get a config from a vega to show a prime example as the app creates the configs itself... so give it a shot;  assuming I understood your initial question properly

Also;  if you add more GPU's to the system, or remove some, or change the CPU, delete the appropriate file (amd.txt, cpu.txt or nvidia.txt) and re-run xmr-stak.exe.  it will re-assess the system and rebuild the configs for the current hardware set.

thnx,
 the vega  can run two simultaneous threads reaching 2000H/s but need a double string input in the amd.txt and for the life of me I cant get the coding correct (spacing? a comma? idunno) I'm sure I will get right by the end of the day by trial and error
I will come back and post it here

[JaredKaragen]

thnx,
 the vega  can run two simultaneous threads reaching 2000H/s but need a double string input in the amd.txt and for the life of me I cant get the coding correct (spacing? a comma? idunno) I'm sure I will get right by the end of the day by trial and error
I will come back and post it here

When you do;  generate a fresh amd.txt by deleting it and running xmr-stak again.  Post it here with the #code posting ability

Code:

like this

Ill have a peek and give some input =)

Ill be able to put the Rx560 I have back on xmr-stak off my other tests to generate a amd.txt and see what i see on my end as well.  Im pretty sure the vega driver supports the version of openCL I compiled it with.

[JaredKaragen]

[cuda] error gpu 0 something something :365

1080ti on windows 10 64 bit. 21gb virtual memory

edit nvidia.txt

see notes at beginning of file about adjusting bfactor and bsleep.  probably happening cause you have a monitor plugged into GPU #0.

[Xapodat]

This miner does not work so well on old cards like HD 7970. I only get about 400H/s. Anybody has similar result?

[RIGED]

thnx,
 the vega  can run two simultaneous threads reaching 2000H/s but need a double string input in the amd.txt and for the life of me I cant get the coding correct (spacing? a comma? idunno) I'm sure I will get right by the end of the day by trial and error
I will come back and post it here

When you do;  generate a fresh amd.txt by deleting it and running xmr-stak again.  Post it here with the #code posting ability

Code:

like this

Ill have a peek and give some input =)

Ill be able to put the Rx560 I have back on xmr-stak off my other tests to generate a amd.txt and see what i see on my end as well.  Im pretty sure the vega driver supports the version of openCL I compiled it with.

got it work with max intensity set at 1012
but only one thread

Code:

"gpu_threads_conf" : [
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 0,
    "intensity" : 1012, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 1,
    "intensity" : 1012, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 2,
    "intensity" : 1012, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 3,
    "intensity" : 1012, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 4,
    "intensity" : 1012, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 5,
    "intensity" : 1012, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },

],

/*
 * Platform index. This will be 0 unless you have different OpenCL platform - eg. AMD and Intel.
 */
"platform_index" : 1,

giving about the same hash results as the best I can get it with a double thread per card (yea it was the intensities that were giving the buffer error message)

Code:

"gpu_threads_conf" : [
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 0,
    "intensity" : 1000, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  { "index" : 0,
    "intensity" : 800, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 1,
    "intensity" : 1000, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  { "index" : 1,
    "intensity" : 800, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 2,
    "intensity" : 1000, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  { "index" : 2,
    "intensity" : 800, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 3,
    "intensity" : 1000, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  { "index" : 3,
    "intensity" : 800, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 4,
    "intensity" : 1000, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  { "index" : 4,
    "intensity" : 800, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  // gpu: gfx901 memory:3920
  // compute units: 64
  { "index" : 5,
    "intensity" : 1000, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },
  { "index" : 5,
    "intensity" : 800, "worksize" : 8,
    "affine_to_cpu" : false, "strided_index" : 1, "mem_chunk" : 2,
    "comp_mode" : true
  },

],

/*
 * Platform index. This will be 0 unless you have different OpenCL platform - eg. AMD and Intel.
 */
"platform_index" : 1,

still playing around with the intensities on the dual thread amd.txt file but it seems the new heavy algo is very demanding on the GPUs as i cant get anywhere near the previous algo 2000H/s

good to be on Cryptonight again

many thanx JK!!