8.16 or Commits Fail to Compile on Linux

[joesixpack]

8.15 is fine. What a pity as I wanted the neoscrypt fix.

[natewalck]

It compiled on CentOS 7 with Cuda 9.1 for me. Instead of a picture, would it be possible to capture the entire output and put it into a gist?

[joesixpack]

Is this what you wanted?

https://gist.github.com/joesixpack/83be52a48f464b02168a15d79c53ca98

In the meantime, I was able to put in the neoscrypt hack into 8.15. No idea yet if it works or not.

[natewalck]

Which cuda version, driver version and distro?

[ksze]

I also got "multiple definition of `ROTL64(unsigned long, unsigned char)'"

Ubuntu 16.04.3 64-bit Server Edition with all updates.

NVIDIA graphics driver and CUDA 9.1.85-1 from NVIDIA's deb repo (deb http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 /):

$ dpkg -l '*nvidia*' | grep '^ii'
ii  nvidia-387                       387.26-0ubuntu1 amd64        NVIDIA binary driver - version 387.26
ii  nvidia-387-dev                   387.26-0ubuntu1 amd64        NVIDIA binary Xorg driver development files
ii  nvidia-modprobe                  387.26-0ubuntu1 amd64        Load the NVIDIA kernel driver and create device files
ii  nvidia-opencl-icd-387            387.26-0ubuntu1 amd64        NVIDIA OpenCL ICD
ii  nvidia-prime                     0.8.2           amd64        Tools to enable NVIDIA's Prime
ii  nvidia-settings                  387.26-0ubuntu1 amd64        Tool for configuring the NVIDIA graphics driver
$ dpkg -l cuda*9*
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name                                         Version                     Architecture                Description
+++-============================================-===========================-===========================-==============================================================================================
ii  cuda-9-1                                     9.1.85-1                    amd64                       CUDA 9.1 meta-package
ii  cuda-command-line-tools-9-1                  9.1.85-1                    amd64                       CUDA command-line tools
ii  cuda-compiler-9-1                            9.1.85-1                    amd64                       CUDA compiler
ii  cuda-cublas-9-1                              9.1.85-1                    amd64                       CUBLAS native runtime libraries
ii  cuda-cublas-dev-9-1                          9.1.85-1                    amd64                       CUBLAS native dev links, headers
ii  cuda-cudart-9-1                              9.1.85-1                    amd64                       CUDA Runtime native Libraries
ii  cuda-cudart-dev-9-1                          9.1.85-1                    amd64                       CUDA Runtime native dev links, headers
ii  cuda-cufft-9-1                               9.1.85-1                    amd64                       CUFFT native runtime libraries
ii  cuda-cufft-dev-9-1                           9.1.85-1                    amd64                       CUFFT native dev links, headers
ii  cuda-cuobjdump-9-1                           9.1.85-1                    amd64                       CUDA cuobjdump
ii  cuda-cupti-9-1                               9.1.85-1                    amd64                       CUDA profiling tools interface.
ii  cuda-curand-9-1                              9.1.85-1                    amd64                       CURAND native runtime libraries
ii  cuda-curand-dev-9-1                          9.1.85-1                    amd64                       CURAND native dev links, headers
ii  cuda-cusolver-9-1                            9.1.85-1                    amd64                       CUDA solver native runtime libraries
ii  cuda-cusolver-dev-9-1                        9.1.85-1                    amd64                       CUDA solver native dev links, headers
ii  cuda-cusparse-9-1                            9.1.85-1                    amd64                       CUSPARSE native runtime libraries
ii  cuda-cusparse-dev-9-1                        9.1.85-1                    amd64                       CUSPARSE native dev links, headers
ii  cuda-demo-suite-9-1                          9.1.85-1                    amd64                       Demo suite for CUDA
ii  cuda-documentation-9-1                       9.1.85-1                    amd64                       CUDA documentation
ii  cuda-driver-dev-9-1                          9.1.85-1                    amd64                       CUDA Driver native dev stub library
ii  cuda-gdb-9-1                                 9.1.85-1                    amd64                       CUDA-GDB
ii  cuda-gpu-library-advisor-9-1                 9.1.85-1                    amd64                       CUDA GPU Library Advisor.
ii  cuda-libraries-9-1                           9.1.85-1                    amd64                       CUDA Libraries 9.1 meta-package
ii  cuda-libraries-dev-9-1                       9.1.85-1                    amd64                       CUDA Libraries 9.1 development meta-package
ii  cuda-license-9-1                             9.1.85-1                    amd64                       CUDA licenses
ii  cuda-memcheck-9-1                            9.1.85-1                    amd64                       CUDA-MEMCHECK
ii  cuda-misc-headers-9-1                        9.1.85-1                    amd64                       CUDA miscellaneous headers
ii  cuda-npp-9-1                                 9.1.85-1                    amd64                       NPP native runtime libraries
ii  cuda-npp-dev-9-1                             9.1.85-1                    amd64                       NPP native dev links, headers
ii  cuda-nsight-9-1                              9.1.85-1                    amd64                       CUDA nsight
ii  cuda-nvcc-9-1                                9.1.85-1                    amd64                       CUDA nvcc
ii  cuda-nvdisasm-9-1                            9.1.85-1                    amd64                       CUDA disassembler
ii  cuda-nvgraph-9-1                             9.1.85-1                    amd64                       NVGRAPH native runtime libraries
ii  cuda-nvgraph-dev-9-1                         9.1.85-1                    amd64                       NVGRAPH native dev links, headers
ii  cuda-nvml-dev-9-1                            9.1.85-1                    amd64                       NVML native dev links, headers
ii  cuda-nvprof-9-1                              9.1.85-1                    amd64                       CUDA Profiler tools
ii  cuda-nvprune-9-1                             9.1.85-1                    amd64                       CUDA nvprune
ii  cuda-nvrtc-9-1                               9.1.85-1                    amd64                       NVRTC native runtime libraries
ii  cuda-nvrtc-dev-9-1                           9.1.85-1                    amd64                       NVRTC native dev links, headers
ii  cuda-nvtx-9-1                                9.1.85-1                    amd64                       NVIDIA Tools Extension
ii  cuda-nvvp-9-1                                9.1.85-1                    amd64                       CUDA nvvp
ii  cuda-runtime-9-1                             9.1.85-1                    amd64                       CUDA Runtime 9.1 meta-package
ii  cuda-samples-9-1                             9.1.85-1                    amd64                       CUDA example applications
ii  cuda-toolkit-9-1                             9.1.85-1                    amd64                       CUDA Toolkit 9.1 meta-package
ii  cuda-tools-9-1                               9.1.85-1                    amd64                       CUDA Tools meta-package
ii  cuda-visual-tools-9-1                        9.1.85-1                    amd64                       CUDA visual tools

[KlausT]

There are no multiple definitions of ROTL64. I didn't touch the ROTL64 stuff since, like, forever. It's a compiler issue, or you are mixing files from different versions.

[drpoom]

I got similar errors like above (collect2, and multiple ROTL64 definitions). Compiled on Ubuntu 16.4, 64-bit desktop. Neither CUDA 9.1 nor 9.0 succeeded, it's probably some other dependencies.

[KlausT]

Let me guess. You are all not using the build.sh script.

By the way, I fixed a small typo in configure.sh

[drpoom]

You are dead right! I used ./autogen.sh, then ./configure.sh, then make. facepalm After using ./build.sh the problem is solved.

[ksze]

I did use the build.sh script and I got that error.

[joesixpack]

No problems compiling now. Guess it was that small typo.

[AlexShpak]

Same here. I'm using build.sh and still getting multiple definition of `ROTL64(unsigned long, unsigned char)' CentOS 7, CUDA 9.1 Driver 384.98

UPD: Sorry, windows branch compiled ok. all above concerns cuda9 branch. And yeah, i know that i should install a newer driver :)

[joesixpack]

Spoke too soon. Compiles on Ubuntu 16.04 LTS but not 17.10.

[KlausT]

Wierd stuff is happening here...

CUDA 8 system requirements:

CUDA 9.1 system requirements:

[eruditej]

I am testing on Ubuntu 16 (history has taught me to never use latest). Things I have learned.

1. run autogen.sh
2. install gcc-4.9 and g++-4.9
3. CC=gcc-4.9 CPP=cpp-4.9 CXX=g++-4.9 ./configure.sh

The gcc 5.4 in current ubuntu 16.04 doesnt work with cuda 9.1. Might be a hard coded invariant in the headers (nvidia loves to do this, even on windows).

It builds, it doesn't work for me but it builds. I am trying to add some unified memory support code for Linux so it is much faster than Windows where Unified memory can't be used and the code spends all it's time run memCpy.

[eruditej]

./ccminer -i21 -a neoscrypt --url=stratum+tcp://neoscrypt.usa.nicehash.com:3341 -u 1PVw8bu 17HWPhLoqQShu5cS5qaoTQRDwFA.titans -p x ccminer 8.17-KlausT (64bit) for nVidia GPUs Compiled with GCC 4.9 using Nvidia CUDA Toolkit 9.1

Based on pooler cpuminer 2.3.2 and the tpruvot@github fork CUDA support by Christian Buchner, Christian H. and DJM34 Includes optimizations implemented by sp-hash, klaust, tpruvot and tsiv.

[2017-12-25 10:31:02] NVML GPU monitoring enabled. [2017-12-25 10:31:02] Intensity set to 21, 2097152 cuda threads [2017-12-25 10:31:02] 2 miner threads started, using 'neoscrypt' algorithm. 0 1 [2017-12-25 10:31:02] Starting Stratum on stratum+tcp://neoscrypt.usa.nicehash.com:3341 [2017-12-25 10:31:02] Stratum difficulty set to 1024 [2017-12-25 10:31:02] neoscrypt.usa.nicehash.com:3341 neoscrypt block 2560 Cuda error in func 'neoscrypt_cpu_init_2stream' at line 1432 : out of memory. Cuda error in func 'scanhash_neoscrypt' at line 41 : driver shutting down. Segmentation fault (core dumped)

So it will build on 16.04, but lyra2v2 and neoscrypt neither work. tpruvot 2.2.3-branch cuda9 does work currently (and has to be built the same way).

I an not familiar enough with the code to say what is going on, but this block is screaming for UMA support:

void neoscrypt_cpu_init_2stream(int thr_id, uint32_t threads)
{
        uint32_t *hash1;
        uint32_t *hash2; // 2 streams
        uint32_t *Trans1;
        uint32_t *Trans2; // 2 streams
        uint32_t *Trans3; // 2 streams
        uint32_t *Bhash;

        CUDA_SAFE_CALL(cudaStreamCreate(&stream[0]));
        CUDA_SAFE_CALL(cudaStreamCreate(&stream[1]));

        CUDA_SAFE_CALL(cudaMalloc(&d_NNonce[thr_id], 2 * sizeof(uint32_t)));
        CUDA_SAFE_CALL(cudaMalloc(&hash1, 32ULL * 128 * sizeof(uint64_t) * threads));
        CUDA_SAFE_CALL(cudaMalloc(&hash2, 32ULL * 128 * sizeof(uint64_t) * threads));
        CUDA_SAFE_CALL(cudaMalloc(&Trans1, 32ULL * sizeof(uint64_t) * threads));
        CUDA_SAFE_CALL(cudaMalloc(&Trans2, 32ULL * sizeof(uint64_t) * threads));
        CUDA_SAFE_CALL(cudaMalloc(&Trans3, 32ULL * sizeof(uint64_t) * threads));
        CUDA_SAFE_CALL(cudaMalloc(&Bhash, 128ULL * sizeof(uint32_t) * threads));

        CUDA_SAFE_CALL(cudaMemcpyToSymbolAsync(B2, &Bhash, sizeof(uint28*), 0, cudaMemcpyHostToDevice, stream[0]));
        CUDA_SAFE_CALL(cudaMemcpyToSymbolAsync(W, &hash1, sizeof(uint28*), 0, cudaMemcpyHostToDevice, stream[0]));
        CUDA_SAFE_CALL(cudaMemcpyToSymbolAsync(W2, &hash2, sizeof(uint28*), 0, cudaMemcpyHostToDevice, stream[0]));
        CUDA_SAFE_CALL(cudaMemcpyToSymbolAsync(Tr, &Trans1, sizeof(uint28*), 0, cudaMemcpyHostToDevice, stream[0]));
        CUDA_SAFE_CALL(cudaMemcpyToSymbolAsync(Tr2, &Trans2, sizeof(uint28*), 0, cudaMemcpyHostToDevice, stream[0]));
        CUDA_SAFE_CALL(cudaMemcpyToSymbolAsync(Input, &Trans3, sizeof(uint28*), 0, cudaMemcpyHostToDevice, stream[0]));
        if(opt_debug)
                CUDA_SAFE_CALL(cudaDeviceSynchronize());

Apparently gcc5.4+ and cuda have known warts:
Differences in compiler macros between Ubuntu 16.04 GCC 5 and Ubuntu 17.10 GCC 5:

ref: https://aur.archlinux.org/cgit/aur.git/commit/?h=ccminer-git&id=0dc0253c4cee236eca1289aafae8fdce7d81fac5

ref2: https://github.com/caffe2/caffe2/issues/1633

TL;DR: its not KlausT (or tpruvot). If you want neoscrypt now though, use trpruvot until the big is fixed causing OOM crash.  

[eruditej]

https://bugs.launchpad.net/ubuntu/+source/gcc-5/+bug/1725848

They went from 5.3 to 5.4, broke the bbuild, wontfix. #justmaintainerthings

[KlausT]

You are still using this wierd source code?

[2017-12-25 10:31:02] 2 miner threads started, using 'neoscrypt' algorithm.
0
1
[2017-12-25 10:31:02] Starting Stratum on stratum+tcp://neoscrypt.usa.nicehash.com:3341

Please try the latest commit from the cuda9 branch.

Also, -i 21 ? That's too high for neoscrypt. What cards do you have?

[eruditej]

dual titan V :-)

I was just testing neoscript for the guy who created the issue.

To sum up: cuda 9.1 works on Ubuntu 16 only w/ gcc 4.9 currently (configuring with gcc-4.9 as the compiler) cuda 9.1 works on Ubuntu 17.04 w/ gcc-5 cuda 9.1 works on Ubuntu 17.10 maybe... using gcc-6

I think this persons' issue can be closed, although they will not like the answer. It isn't the code, its the toolchain.

[eruditej]

Updated my rig to ubuntu 17.04 (which might be the current sweet spot for cuda9.1)

my configure.sh extracflags="-march=native -D_REENTRANT -falign-functions=16 -falign-jumps=16 -falign-labels=16"

CUDA_CFLAGS='-lineno --shared --compiler-options "-Wall -fPIC"' \ ./configure CXXFLAGS="-O3 $extracflags" --with-cuda=/usr/local/cuda --with-nvml=libnvidia-ml.so

Ultimately, people are going to have to learn to use autoconfig themselves versus expecting a guide or wrapper that works for all the toolchain combos.

Result:

./ccminer -a neoscrypt --url=stratum+tcp://neoscrypt.usa.nicehash.com:3341 -u .titans -p x ccminer 8.17-KlausT (64bit) for nVidia GPUs Compiled with GCC 6.3 using Nvidia CUDA Toolkit 9.1

Based on pooler cpuminer 2.3.2 and the tpruvot@github fork CUDA support by Christian Buchner, Christian H. and DJM34 Includes optimizations implemented by sp-hash, klaust, tpruvot and tsiv.

[2017-12-25 13:10:18] NVML GPU monitoring enabled. 0 1 [2017-12-25 13:10:18] Starting Stratum on stratum+tcp://neoscrypt.usa.nicehash.com:3341 [2017-12-25 13:10:18] 2 miner threads started, using 'neoscrypt' algorithm. [2017-12-25 13:10:18] Stratum difficulty set to 256 [2017-12-25 13:10:18] neoscrypt.usa.nicehash.com:3341 neoscrypt block 2637 [2017-12-25 13:10:22] accepted: 1/1 (100.00%), 998.13 kH/s yay!!! [2017-12-25 13:10:33] GPU #1: Graphics Device, 1398.44 kH/s [2017-12-25 13:10:33] accepted: 2/2 (100.00%), 2596.48 kH/s yay!!! [2017-12-25 13:10:33] GPU #1: Graphics Device, 1413.00 kH/s [2017-12-25 13:10:34] accepted: 3/3 (100.00%), 2611.04 kH/s yay!!! [2017-12-25 13:10:44] GPU #0: Graphics Device, 1440.68 kH/s [2017-12-25 13:10:48] GPU #0: Graphics Device, 1463.33 kH/s [2017-12-25 13:10:48] accepted: 4/4 (100.00%), 2865.01 kH/s yay!!! [2017-12-25 13:10:49] neoscrypt.usa.nicehash.com:3341 neoscrypt block 2637 [2017-12-25 13:10:49] GPU #0: Graphics Device, 1477.40 kH/s [2017-12-25 13:10:49] GPU #1: Graphics Device, 1426.44 kH/s [2017-12-25 13:10:51] GPU #0: Graphics Device, 1418.96 kH/s [2017-12-25 13:10:51] accepted: 5/5 (100.00%), 2869.81 kH/s yay!!! [2017-12-25 13:10:52] GPU #0: Graphics Device, 1443.67 kH/s [2017-12-25 13:10:52] accepted: 6/6 (100.00%), 2868.53 kH/s yay!!!

unfortunately: ./ccminer -a lyra2v2 --benchmark ccminer 8.17-KlausT (64bit) for nVidia GPUs Compiled with GCC 6.3 using Nvidia CUDA Toolkit 9.1

Based on pooler cpuminer 2.3.2 and the tpruvot@github fork CUDA support by Christian Buchner, Christian H. and DJM34 Includes optimizations implemented by sp-hash, klaust, tpruvot and tsiv.

[2017-12-25 13:17:14] NVML GPU monitoring enabled. 0 [2017-12-25 13:17:14] 2 miner threads started, using 'lyra2v2' algorithm. 1 [2017-12-25 13:17:15] GPU #1: result does not validate on CPU! [2017-12-25 13:17:15] GPU #1: result does not validate on CPU! [2017-12-25 13:17:15] GPU #0: result does not validate on CPU!

[KlausT]

I think that you have old and new files mixed together.

[eruditej]

I'm using current cuda9 branch. Git status shows I am up to date except for my own modified configure.sh, and I've run make clean.

[KlausT]

"0" and "1" doesn't appear in the current version. It's probably a leftover from an old version where I had added that for debugging purposes.

[eruditej]

https://github.com/KlausT/ccminer/blob/ef144cf7ffa4073122c2b128a6622b2cf32b8ea7/lyra2/lyra2REv2.cu#L217

cuda9 branch

[KlausT]

Yeah, it should be device_map[thr_id] But this doesn't explain the "0" and "1"

[2017-12-25 13:10:18] NVML GPU monitoring enabled.
0
1
[2017-12-25 13:10:18] Starting Stratum on stratum+tcp://neoscrypt.usa.nicehash.com:3341

[eruditej]

jeramy@miner1:~/repos/KlausT$ nvidia-smi
Mon Dec 25 14:09:34 2017
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 387.34                 Driver Version: 387.34                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Graphics Device     Off  | 00000000:01:00.0 Off |                  N/A |
| 54%   73C    P2   141W / 250W |   3579MiB / 12055MiB |     97%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Graphics Device     Off  | 00000000:03:00.0 Off |                  N/A |
| 49%   68C    P2   145W / 250W |   3579MiB / 12058MiB |     97%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      9076      C   ./ccminer                                   3568MiB |
|    1      9076      C   ./ccminer                                   3568MiB |
+-----------------------------------------------------------------------------+

nvidia linux-ism?

[eruditej]

Nevermind, upgrade killed my driver. Hence, it didnt show up as Titan, fixing locally.

[KlausT]

Would you like to do a little experiment? You might be surprised. Download this: https://github.com/KlausT/ccminer/archive/cuda9.zip then run build.sh and try to mine lyra2v2.

[eruditej]

ref: https://devtalk.nvidia.com/default/topic/1027645/linux/nvidia-smi-not-recognizing-titan-v-/post/5228400/#5228400

"Graphics Device" showing up for Volta Titans on Linux is a known issue.

I added a few lines to my local Lyra2Rev to add default intensity when the card is "too new and awesome". Testing your build now.

fyi: nvidia-smi -L GPU 0: Graphics Device (UUID: GPU-e3970f09-7b4b-59f7-7281-05cd4be92e74) GPU 1: Graphics Device (UUID: GPU-48c1bc72-b872-29a3-cbd5-dcf835f74a14)

thats w/ latest drivers/cuda on linux, fun!

[KlausT]

Maybe I should check for the string "Graphics Device" and assume it's a Titan V :-)

[eruditej]

stratum still fails with my local and your test,

this doesnt fix anything, but it makes it clear its working as intended

diff --git a/lyra2/lyra2REv2.cu b/lyra2/lyra2REv2.cu
index 7663a29..d507fcb 100644
--- a/lyra2/lyra2REv2.cu
+++ b/lyra2/lyra2REv2.cu
@@ -83,7 +83,15 @@ int scanhash_lyra2v2(int thr_id, uint32_t *pdata,

        cudaDeviceProp props;
        cudaGetDeviceProperties(&props, device_map[thr_id]);
-       if(strstr(props.name, "Titan"))
+       // if its too new and awesome, linux will give it a generic name
+       if(strstr(props.name, "Graphics Device"))
+       {
+               intensity = 256 * 256 * 15;
+#ifdef _WIN64
+               intensity = 256 * 256 * 22;
+#endif
+       }
+       else if(strstr(props.name, "Titan"))
        {
                intensity = 256 * 256 * 15;
 #ifdef _WIN64

[KlausT]

Oh, and I have to change the ifdef After all this time I still find silly bugs.

[eruditej]

I am getting connected to the stratum, but still not validating. tpruvot rebuilt on the same chain working.

./ccminer -i21 -a lyra2v2 --url=stratum+tcp://lyra2rev2.usa.nicehash.com:3347 -u <redacted>.titans -p x
ccminer 8.17-KlausT (64bit) for nVidia GPUs
Compiled with GCC 6.3 using Nvidia CUDA Toolkit 9.1

Based on pooler cpuminer 2.3.2 and the tpruvot@github fork
CUDA support by Christian Buchner, Christian H. and DJM34
Includes optimizations implemented by sp-hash, klaust, tpruvot and tsiv.

[2017-12-25 15:36:25] NVML GPU monitoring enabled.
[2017-12-25 15:36:25] Intensity set to 21, 2097152 cuda threads
0
1
[2017-12-25 15:36:25] 2 miner threads started, using 'lyra2v2' algorithm.
[2017-12-25 15:36:25] Starting Stratum on stratum+tcp://lyra2rev2.usa.nicehash.com:3347
[2017-12-25 15:36:25] Stratum difficulty set to 32
[2017-12-25 15:36:25] lyra2rev2.usa.nicehash.com:3347 lyra2v2 block 1749227
[2017-12-25 15:36:30] GPU #0: result does not validate on CPU!
[2017-12-25 15:36:30] GPU #0: result does not validate on CPU!

[KlausT]

Since it works on Windows, it could be unexpected compiler behaviour or wrong compiler options. Unfortunately I can't help you there.

[KlausT]

try the option --no-cpu-verify then we can see where the problem is, the CUDA code, or the CPU code

[eruditej]

lol, smoking

 ../KlausT/ccminer -i21 -a lyra2v2 --url=stratum+tcp://lyra2rev2.usa.nicehash.com:3347 -u
<redacted>.titans -p x --no-cpu-verify
ccminer 8.17-KlausT (64bit) for nVidia GPUs
Compiled with GCC 6.3 using Nvidia CUDA Toolkit 9.1

Based on pooler cpuminer 2.3.2 and the tpruvot@github fork
CUDA support by Christian Buchner, Christian H. and DJM34
Includes optimizations implemented by sp-hash, klaust, tpruvot and tsiv.

[2017-12-25 16:30:29] NVML GPU monitoring enabled.
[2017-12-25 16:30:29] Intensity set to 21, 2097152 cuda threads
0
1
[2017-12-25 16:30:29] Starting Stratum on stratum+tcp://lyra2rev2.usa.nicehash.com:3347
[2017-12-25 16:30:29] 2 miner threads started, using 'lyra2v2' algorithm.
[2017-12-25 16:30:29] Stratum difficulty set to 32
[2017-12-25 16:30:29] lyra2rev2.usa.nicehash.com:3347 lyra2v2 block 1749316
[2017-12-25 16:30:32] accepted: 1/1 (100.00%), 92.15 MH/s yay!!!
[2017-12-25 16:30:33] accepted: 2/2 (100.00%), 92.15 MH/s yay!!!
[2017-12-25 16:30:33] lyra2rev2.usa.nicehash.com:3347 lyra2v2 block 1749317
[2017-12-25 16:30:33] GPU #1: Graphics Device, 97.87 MH/s
[2017-12-25 16:30:33] GPU #0: Graphics Device, 98.48 MH/s
[2017-12-25 16:30:37] GPU #1: Graphics Device, 99.22 MH/s
[2017-12-25 16:30:37] accepted: 3/3 (100.00%), 197.70 MH/s yay!!!
[2017-12-25 16:30:37] GPU #1: Graphics Device, 99.25 MH/s
[2017-12-25 16:30:37] accepted: 4/4 (100.00%), 197.71 MH/s yay!!!
[2017-12-25 16:30:37] GPU #1: Graphics Device, 99.31 MH/s
[2017-12-25 16:30:37] accepted: 5/5 (100.00%), 197.74 MH/s yay!!!
[2017-12-25 16:30:39] GPU #1: Graphics Device, 99.54 MH/s
[2017-12-25 16:30:39] accepted: 6/6 (100.00%), 197.81 MH/s yay!!!

[cgarnier]

Is it possible to provide a docker image for the different working version ? I m not used to build cpp stuff and i m fighting with it :D I m trying to adapt this one https://github.com/kahiroka/ccminer-in-docker/blob/master/Dockerfile with your fork but for now it s not working.

[KlausT]

Would you please open a seperate issue for the docker stuff? Then other people will actually find it. Thanks

Bye the way, I'm not a Linux user, so I will have to rely on other people who know more about Linux than me.

[sukoshi1507]

Any error message to share ?

Le 22 janv. 2018 17:33, "Clément Garnier" notifications@github.com a écrit :

Is it possible to provide a docker image for the different working version ? I m not used to build cpp stuff and i m fighting with it :D I m trying to adapt this one https://github.com/kahiroka/ ccminer-in-docker/blob/master/Dockerfile with your fork but for now it s not working.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/KlausT/ccminer/issues/92#issuecomment-359482943, or mute the thread https://github.com/notifications/unsubscribe-auth/AQeB2plLo9GkVYucqQb7p1Dizi9Pqz2uks5tNLhXgaJpZM4REzaN .

[eruditej]

Yes, its possible to dock this. I run it that way. It works fine.

The only thing needed is a Dockerfile, which I can add to my tree and ask klaust to pull.

By using Docker, you can have tpruvot running in a xenial-cuda9 docker and klaust running in a xenial-cuda9.1 docker on the same underlying base OS (in my case, Zesty).

It seems under cuda-9.1 on linux, you have to disable cpu verify (which tpruvot does not have an option for, which is why he's still cuda9).

For anyone who finds this later, no, you cannot use Docker for Windows and load a Xenial Ubuntu Docker for Mining passed through to host GPU. https://github.com/NVIDIA/nvidia-docker/issues/428

Most Linux users who mine with KlausT on Linux are hermits who do not share their work. They run farms and make it work for themselves. I will share.

On Mon, Jan 22, 2018 at 10:39 AM, sukoshi1507 notifications@github.com wrote:

Any error message to share ?

Le 22 janv. 2018 17:33, "Clément Garnier" notifications@github.com a écrit :

Is it possible to provide a docker image for the different working version ? I m not used to build cpp stuff and i m fighting with it :D I m trying to adapt this one https://github.com/kahiroka/ ccminer-in-docker/blob/master/Dockerfile with your fork but for now it s not working.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/KlausT/ccminer/issues/92#issuecomment-359482943, or mute the thread https://github.com/notifications/unsubscribe-auth/ AQeB2plLo9GkVYucqQb7p1Dizi9Pqz2uks5tNLhXgaJpZM4REzaN

.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/KlausT/ccminer/issues/92#issuecomment-359485202, or mute the thread https://github.com/notifications/unsubscribe-auth/AhKu4DxNH6r4lON7HfDxgjsGUFcJ8dJwks5tNLm6gaJpZM4REzaN .

[cgarnier]

Could you publish your dockerfiles ?