[OS] nvOC easy-to-use Linux Nvidia Mining

[CryptAtomeTrader44]

I have a problem from several month with xmr-stak :

When xmr-stak start, i can read theses messages :

[2019-03-09 16:31:28] : Mining coin: cryptonight_r
[2019-03-09 16:31:28] : NVIDIA: try to load library 'xmrstak_cuda_backend_cuda10_0'
WARNING: NVIDIA cannot load backend library: libxmrstak_cuda_backend_cuda10_0.so: cannot open shared object file: No such file or directory
[2019-03-09 16:31:28] : NVIDIA: try to load library 'xmrstak_cuda_backend_cuda9_2'
WARNING: NVIDIA cannot load backend library: libxmrstak_cuda_backend_cuda9_2.so: cannot open shared object file: No such file or directory
[2019-03-09 16:31:28] : NVIDIA: try to load library 'xmrstak_cuda_backend'
NVIDIA: found 10 potential device's

but when i use nvidia-smi command, i can see that i have CUDA 10.0 :

nvidia-smi
Sat Mar  9 16:16:40 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.66       Driver Version: 410.66       CUDA Version: 10.0     |

and when a i compiled the latest xmr-stak version, i saw that several backend had been compiled :

[ 70%] Built target xmr-stak-backend
[ 72%] Building NVCC (Device) object CMakeFiles/xmrstak_cuda_backend.dir/xmrstak                                                   /backend/nvidia/nvcc_code/xmrstak_cuda_backend_generated_cuda_extra.cu.o
[ 75%] Building NVCC (Device) object CMakeFiles/xmrstak_cuda_backend.dir/xmrstak                                                   /backend/nvidia/nvcc_code/xmrstak_cuda_backend_generated_cuda_core.cu.o

List of files in my xmr-stak folder :

:~/NVOC/mining/miners/XMR_Stak/2.10.0$ ll
total 50104
drwxrwxr-x 2 m1   m1       4096 Mar  9 16:26 ./
drwxrwxr-x 9 m1   m1       4096 Mar  9 16:30 ../
-rw-rw-r-- 1 m1   m1       6729 Mar  9 16:22 config.txt
-rw-r--r-- 1 root root    58488 Mar  9 16:21 libxmr-stak-asm.a
-rw-r--r-- 1 root root  2569042 Mar  9 16:21 libxmr-stak-backend.a
-rw-r--r-- 1 root root    74734 Mar  9 16:21 libxmr-stak-c.a
-rwxr-xr-x 1 root root 44997624 Mar  9 16:21 libxmrstak_cuda_backend.so*
-rwxr-xr-x 1 root root  2014400 Mar  9 16:21 libxmrstak_opencl_backend.so*
-rw-rw-r-- 1 m1   m1       4606 Mar  9 16:24 nvidia.txt
-rw-rw-r-- 1 m1   m1       2482 Mar  9 16:24 pools.txt
-rwxr-xr-x 1 root root  1550280 Mar  9 16:21 xmr-stak*

What's the problem on my system that indicate i have not the good CUDA backend ?

How can i  remedy/resolve this problem ?

[papampi]

Today I noticed it too, while I have not installed cuda-10 on my rig nvidia-smi shows cuda-10
But my default cuda version is 9.2

Code:

m1@m1-desktop:~$ cat /usr/local/cuda/version.txt
CUDA Version 9.2.148

Have you tried beta_testing: bash nvOC miners-upgrade
Then let it recompile XMR_Stak?

[CryptAtomeTrader44]

Today I noticed it too, while I have not installed cuda-10 on my rig nvidia-smi shows cuda-10
But my default cuda version is 9.2

Code:

m1@m1-desktop:~$ cat /usr/local/cuda/version.txt
CUDA Version 9.2.148

I have the same problem while i don't upgrade yet.

Have you tried beta_testing: bash nvOC miners-upgrade
Then let it recompile XMR_Stak?

I recompiled xmr-stak from 2.9.0 to 2.10.0 at 16h30 like my last message looks.

I will upgrade later. I'll have to test this foreman interface. For sure.



For now, it seems that the fork of monero has not changed anything, it has even made things worse since there is 90% of the hashrate that is dominated by ASICs or AMAPs, against 60-70% before the fork , which makes the network still more vulnerable! AND does not prevent extreme centralization.

[obm_dan]

If you'd like to see how a Foreman dashboard looks, we just released a public read-only demo account that you can access here:  https://dashboard.foreman.mn/demo/

[papampi]

Different cuda versions shown by nvcc and nvidia-smi:

CUDA has 2 primary APIs, the runtime and the driver API. Both have a corresponding version (e.g. 8.0, 9.0, etc.)

The necessary support for the driver API (e.g. libcuda.so on linux) is installed by the GPU driver installer.

The necessary support for the runtime API (e.g. libcudart.so on linux, and also nvcc) is installed by the CUDA toolkit installer (which may also have a GPU driver installer bundled in it).

In any event, the (installed) driver API version may not always match the (installed) runtime API version, especially if you install a GPU driver independently from installing CUDA (i.e. the CUDA toolkit).

The nvidia-smi tool gets installed by the GPU driver installer, and generally has the GPU driver in view, not anything installed by the CUDA toolkit installer.

Recently (somewhere between 410.48 and 410.73 driver version on linux) the powers-that-be at NVIDIA decided to add reporting of the CUDA Driver API version installed by the driver, in the output from nvidia-smi.

This has no connection to the installed CUDA runtime version.

nvcc, the CUDA compiler-driver tool that is installed with the CUDA toolkit, will always report the CUDA runtime version that it was built to recognize. It doesn't know anything about what driver version is installed, or even if a GPU driver is installed.

Therefore, by design, these two numbers don't necessarily match, as they are reflective of two different things.

[CryptAtomeTrader44]

Different cuda versions shown by nvcc and nvidia-smi:

CUDA has 2 primary APIs, the runtime and the driver API. Both have a corresponding version (e.g. 8.0, 9.0, etc.)

The necessary support for the driver API (e.g. libcuda.so on linux) is installed by the GPU driver installer.

The necessary support for the runtime API (e.g. libcudart.so on linux, and also nvcc) is installed by the CUDA toolkit installer (which may also have a GPU driver installer bundled in it).

In any event, the (installed) driver API version may not always match the (installed) runtime API version, especially if you install a GPU driver independently from installing CUDA (i.e. the CUDA toolkit).

The nvidia-smi tool gets installed by the GPU driver installer, and generally has the GPU driver in view, not anything installed by the CUDA toolkit installer.

Recently (somewhere between 410.48 and 410.73 driver version on linux) the powers-that-be at NVIDIA decided to add reporting of the CUDA Driver API version installed by the driver, in the output from nvidia-smi.

This has no connection to the installed CUDA runtime version.

nvcc, the CUDA compiler-driver tool that is installed with the CUDA toolkit, will always report the CUDA runtime version that it was built to recognize. It doesn't know anything about what driver version is installed, or even if a GPU driver is installed.

Therefore, by design, these two numbers don't necessarily match, as they are reflective of two different things.

So, perhaps we have to indicate the path when we compile the program ? In makefile ?

My miner mine, but i don't know which cuda version it's using ...I have find how to determine and do the best compilation of xmr-stak.

[papampi]

Different cuda versions shown by nvcc and nvidia-smi:

CUDA has 2 primary APIs, the runtime and the driver API. Both have a corresponding version (e.g. 8.0, 9.0, etc.)

The necessary support for the driver API (e.g. libcuda.so on linux) is installed by the GPU driver installer.

The necessary support for the runtime API (e.g. libcudart.so on linux, and also nvcc) is installed by the CUDA toolkit installer (which may also have a GPU driver installer bundled in it).

In any event, the (installed) driver API version may not always match the (installed) runtime API version, especially if you install a GPU driver independently from installing CUDA (i.e. the CUDA toolkit).

The nvidia-smi tool gets installed by the GPU driver installer, and generally has the GPU driver in view, not anything installed by the CUDA toolkit installer.

Recently (somewhere between 410.48 and 410.73 driver version on linux) the powers-that-be at NVIDIA decided to add reporting of the CUDA Driver API version installed by the driver, in the output from nvidia-smi.

This has no connection to the installed CUDA runtime version.

nvcc, the CUDA compiler-driver tool that is installed with the CUDA toolkit, will always report the CUDA runtime version that it was built to recognize. It doesn't know anything about what driver version is installed, or even if a GPU driver is installed.

Therefore, by design, these two numbers don't necessarily match, as they are reflective of two different things.

So, perhaps we have to indicate the path when we compile the program ? In makefile ?

My miner mine, but i don't know which cuda version it's using ...I have find how to determine and do the best compilation of xmr-stak.

Default Cuda installed in nvOC is 9.2

I think if you use the pre-compiled version from GitHub, it uses cuda from nvidia 410 driver, which is cuda 10

If recompile it, it should use cuda 9.2 from nvOC


Test both see which one gives better hash rate.

[CryptAtomeTrader44]

Default Cuda installed in nvOC is 9.2

I think if you use the pre-compiled version from GitHub, it uses cuda from nvidia 410 driver, which is cuda 10

If recompile it, it should use cuda 9.2 from nvOC


Test both see which one gives better hash rate.

No, i compiled Xmr-stak 2.10.0 like my "how to" described last saturday.

I manually compiled it after a git clone in src folder of xmr-stak.

[infowire]

How do i use the trex miner for RVN ? i see in in folder but its not in 1bash ?

[Tigel]

How do i use the trex miner for RVN ? i see in in folder but its not in 1bash ?

Change:
X16R_MINER="ZENEMYminer"

to:
X16R_MINER="T_Rex"

--
Tigel

[deskless]

Is RadeonVII supported?

[Tigel]

Is RadeonVII supported?

Nope, it's Nvidia only.

--
Tigel

[infowire]

How do i use the trex miner for RVN ? i see in in folder but its not in 1bash ?

Change:
X16R_MINER="ZENEMYminer"

to:
X16R_MINER="T_Rex"

--
Tigel

Thank you, do you know why i get this :

COIN/ALGO not found, Check your settings
Miner not started, Stopping watchdog

my settings

Code:

COIN="RVN"

# Raven Coin (RVN)
RVN_ADDRESS="infowire.RVNMiner1"
RVN_EXTENSION_ARGUMENTS="-p x -a x16r"
RVN_POOL="rvn.suprnova.cc"
RVN_PORT="6666"
RVN_WORKER="$WORKERNAME"

[papampi]

How do i use the trex miner for RVN ? i see in in folder but its not in 1bash ?

Change:
X16R_MINER="ZENEMYminer"

to:
X16R_MINER="T_Rex"

--
Tigel

Thank you, do you know why i get this :

COIN/ALGO not found, Check your settings
Miner not started, Stopping watchdog

my settings

Code:

COIN="RVN"

# Raven Coin (RVN)
RVN_ADDRESS="infowire.RVNMiner1"
RVN_EXTENSION_ARGUMENTS="-p x -a x16r"
RVN_POOL="rvn.suprnova.cc"
RVN_PORT="6666"
RVN_WORKER="$WORKERNAME"

Which nvOC version are you using?

Is X16R coins set in 1bash?

Code:

X16R_COINS="RVN"

[infowire]

How do i use the trex miner for RVN ? i see in in folder but its not in 1bash ?

Change:
X16R_MINER="ZENEMYminer"

to:
X16R_MINER="T_Rex"

--
Tigel

Thank you, do you know why i get this :

COIN/ALGO not found, Check your settings
Miner not started, Stopping watchdog

my settings

Code:

COIN="RVN"

# Raven Coin (RVN)
RVN_ADDRESS="infowire.RVNMiner1"
RVN_EXTENSION_ARGUMENTS="-p x -a x16r"
RVN_POOL="rvn.suprnova.cc"
RVN_PORT="6666"
RVN_WORKER="$WORKERNAME"

Which nvOC version are you using?

Is X16R coins set in 1bash?

Code:

X16R_COINS="RVN"

Code:

X16R_COINS="RVN,XMN"

nvOC_19-3.x_U16.04_Cuda_8-9.2_N410_2018-10-27

[papampi]

nvOC Community Release is moving to its own thread

Please post your questions, comments, ... in new thread

[OS] nvOC Community Release free-easy-to-use Linux Nvidia Mining