Dc5boost44 commented 7 years ago

Getting a GPU Compute Error under error details when im looking at results.

tabcodes commented 7 years ago

Also getting lots of this. Running 2X GTX970's with the following config:

"gpu_threads_conf" : [ { "index" : 0, "threads" : 14, "blocks" : 36, "bfactor" : 7, "bsleep" : 90, "affine_to_cpu" : true, }, { "index" : 1, "threads" : 16, "blocks" : 39, "bfactor" : 0, "bsleep" : 0, "affine_to_cpu" : true, }, ],

Dc5boost44 commented 7 years ago

i turned down my OC seems like its not so often, still got 159 of them last night

psychocrypt commented 7 years ago

thx for reporting: Which OS do you use?

Dc5boost44 commented 7 years ago

Running windows 10 x64 pro. I stopped getting there error when I made my OC less aggressive

tabcodes commented 7 years ago

Linux Mint 18.1 (x64)- turned off OC'ing completely, still getting the error. Changed the first GPU's bfactor/bsleep to 0/0, will report if the error is still occurring when I have access to the machine. Thanks!

larry-a4 commented 7 years ago

Linux 16.04: Getting 1 error every 3 min or so. Overclock.net claims cause to be Nvidia driver. Cannot confirm. http://www.overclock.net/t/1212649/computation-error-on-all-gpu-tasks

tabcodes commented 7 years ago

About the same- I'm running 384.69 and getting the GPU compute error roughly 40-50% of the time. Going to downgrade to 375.82 and see how that goes.

EDIT: Same issue on 375.82.

codingoyster commented 7 years ago

Same issue here! Windows 10 x64, 384.69. Tried increasing power and lowering overclock, no luck. Getting lots of this over time before the .exe crashes on its own. Takes a few days, but still crashes.

jjrev commented 7 years ago

Same issue w/ GTX 980, seeing ~40% error rates, running Ubuntu 16.04, 375.66 drivers, CUDA 7.5, no OC, ~70C@112W

psychocrypt commented 7 years ago

Please post the type of the error. Post the output after you pressed key 'r'. There is somewhere a bug that errors get count twice, we are currently searching where the error is comming from.

Am 01.10.2017 19:07 schrieb "jjrev" notifications@github.com:

Same issue w/ GTX 980, seeing ~40% error rates, running Ubuntu 16.04, 375.66 drivers, CUDA 7.5, no OC, ~70C@112W

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/fireice-uk/xmr-stak-nvidia/issues/124#issuecomment-333391180, or mute the thread https://github.com/notifications/unsubscribe-auth/AYsxtiT7P0EmkuUGrpHOwEG9u3tJ-pOwks5sn8bjgaJpZM4PSK8v .

codingoyster commented 7 years ago

Only error I see is this, psychocrypt:

[GPU COMPUTE ERROR]

3 | 2017-10-01 03:29:27

I have been lowering my overclocks, but still seeing the errors, with xmrstak crashing eventually.

psychocrypt commented 7 years ago

Which pool do you use?

Am 01.10.2017 7:21 Nachm. schrieb "codingoyster" notifications@github.com:

Only error I see is this, psychocrypt: [GPU COMPUTE ERROR]

3 | 2017-10-01 03:29:27

I have been lowering my overclocks, but still seeing the errors, with xmrstak crashing eventually.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/fireice-uk/xmr-stak-nvidia/issues/124#issuecomment-333392045, or mute the thread https://github.com/notifications/unsubscribe-auth/AYsxtgRaRMPNmBEN_ZzXizGarcTJKS9nks5sn8ovgaJpZM4PSK8v .

codingoyster commented 7 years ago

mine.moneropool.com

codingoyster commented 7 years ago

BTW, I also mine on a different machine using xmr-stak-amd. Full overclock, but very rarely see the GPU Compute error. So definitely an issue only with xmr-stak-nvidia.

psychocrypt commented 7 years ago

how long does it take until the first errors will be produced. Could you post the log file with full log level enabled. Currently I don't know if it is a system error or a miner bug. Any time I test xmr-stak-nvidia everything works fine.

Am 01.10.2017 7:52 Nachm. schrieb "codingoyster" notifications@github.com:

BTW, I also mine on a different machine using xmr-stak-amd. Full overclock, but very rarely see the GPU Compute error. So definitely an issue only with xmr-stak-nvidia.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/fireice-uk/xmr-stak-nvidia/issues/124#issuecomment-333393994, or mute the thread https://github.com/notifications/unsubscribe-auth/AYsxtua8n3T_J9D2kNV13d9dFhxXtKWvks5sn9FLgaJpZM4PSK8v .

codingoyster commented 7 years ago

The errors appear early on (within 10-30 mins), but randomly. Sometimes the count has gone >75 before crashing and others within 30-40. I can say that in mining with other ccminer forks, I have never run into GPU errors. I'll enable full log next time I relaunch and share.

jjrev commented 7 years ago

usxmrpool 'r':

RESULT REPORT
Difficulty       : 9602
Good results     : 5 / 8 (62.5 %)
Avg result time  : 38.0 sec
Pool-side hashes : 39324

Top 10 best results found:
|  0 |            25922 |  1 |            18609 |
|  2 |             7967 |  3 |             7044 |
|  4 |             6681 |  5 |                0 |
|  6 |                0 |  7 |                0 |
|  8 |                0 |  9 |                0 |

Error details:
| Count | Error text                       | Last seen           |
|     3 | [GPU COMPUTE ERROR]              | 2017-10-01 11:34:40 |

frankcoderL commented 7 years ago

Hi, I can reproduce the problem and it happens ALL the time with xmr-stak-nvidia. CPU and AMD version are not affected and do work well. Not a single good result, only gpu compute errors. Problem occurs also on aeon-stak-nvidia. The hashrate that is displayed is about 10% higher as usual, but only gets gpu compute errors.

I suppose a recent commit have caused it? A compilation of 28/07/2017 is not affected with this problem.

Is it possible to produce more debug output that can help you solve this issue? Than i will provide it.

Please try Arch Linux to investigate this problem. Latest version of xmr-stak-nvidia and aeon-stak-nvidia seems not to work at all anymore on Arch Linux. Thanks!

[2017-10-05 21:41:42] : Connected. Logging in...
[2017-10-05 21:41:42] : Difficulty changed. Now: 10000.
[2017-10-05 21:41:42] : New block detected.
HASHRATE REPORT
| ID |  10s |  60s |  15m |
|  0 | (na) | (na) | (na) |
---------------------------
Totals:   (na) (na) (na) H/s
Highest:  0.0 H/s
RESULT REPORT
Difficulty       : 10000
Good results     : 0 / 1 (0.0 %)
Pool-side hashes : 0

Top 10 best results found:
|  0 |                0 |  1 |                0 |
|  2 |                0 |  3 |                0 |
|  4 |                0 |  5 |                0 |
|  6 |                0 |  7 |                0 |
|  8 |                0 |  9 |                0 |

Error details:
| Count | Error text                       | Last seen           |
|     1 | [GPU COMPUTE ERROR]              | 2017-10-05 21:41:44 |

HASHRATE REPORT
| ID |  10s |  60s |  15m |
|  0 | 317.5 | 317.7 | (na) |
---------------------------
Totals:   317.5 317.7 (na) H/s
Highest:  317.8 H/s
RESULT REPORT
Difficulty       : 10000
Good results     : 0 / 3 (0.0 %)
Pool-side hashes : 0

Top 10 best results found:
|  0 |                0 |  1 |                0 |
|  2 |                0 |  3 |                0 |
|  4 |                0 |  5 |                0 |
|  6 |                0 |  7 |                0 |
|  8 |                0 |  9 |                0 |

Error details:
| Count | Error text                       | Last seen           |
|     3 | [GPU COMPUTE ERROR]              | 2017-10-05 21:43:33 |

FYI

$ git clone https://github.com/fireice-uk/xmr-stak-nvidia
Cloning into 'xmr-stak-nvidia'...
remote: Counting objects: 353, done.
remote: Total 353 (delta 0), reused 0 (delta 0), pack-reused 353
Receiving objects: 100% (353/353), 342.69 KiB | 1.04 MiB/s, done.
Resolving deltas: 100% (175/175), done.
$ cd xmr-stak-nvidia/
$ export CC=/usr/bin/clang
$ export CXX=/usr/bin/clang++
$ export LDFLAGS=-L/opt/cuda/lib64
$ cmake . -DCUDA_COMPILER=clang -DMICROHTTPD_REQUIRED=OFF
-- The C compiler identification is Clang 5.0.0
-- The CXX compiler identification is Clang 5.0.0
-- Check for working C compiler: /usr/bin/clang
-- Check for working C compiler: /usr/bin/clang -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/clang++
-- Check for working CXX compiler: /usr/bin/clang++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Found CUDA: /opt/cuda (found suitable version "8.0", minimum required is "7.5") 
-- Found OpenSSL: /usr/lib/libcrypto.so (found version "1.1.0f") 
-- Configuring done
-- Generating done
-- Build files have been written to: /home/frank/source/nvidia/mon_new/xmr-stak-nvidia
$ make
Scanning dependencies of target xmr-stak-nvidiaCrypto
[  4%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_blake256.c.o
[  9%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_groestl.c.o
[ 14%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_jh.c.o
[ 19%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_keccak.c.o
[ 23%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_skein.c.o
[ 28%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/soft_aes.c.o
[ 33%] Linking C static library libxmr-stak-nvidiaCrypto.a
[ 33%] Built target xmr-stak-nvidiaCrypto
Scanning dependencies of target xmr-stak-nvidiaCuda
[ 38%] Building CXX object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o
[ 42%] Building CXX object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_extra.cu.o
[ 47%] Linking CXX static library libxmr-stak-nvidiaCuda.a
[ 47%] Built target xmr-stak-nvidiaCuda
Scanning dependencies of target xmr-stak-nvidia
[ 52%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/cli-miner.cpp.o
[ 57%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/console.cpp.o
[ 61%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/executor.cpp.o
[ 66%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/httpd.cpp.o
[ 71%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/jconf.cpp.o
[ 76%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/jpsock.cpp.o
[ 80%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/minethd.cpp.o
[ 85%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/socket.cpp.o
[ 90%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/webdesign.cpp.o
[ 95%] Building CXX object CMakeFiles/xmr-stak-nvidia.dir/crypto/cryptonight_common.cpp.o
[100%] Linking CXX executable bin/xmr-stak-nvidia
[100%] Built target xmr-stak-nvidia
$

ghost commented 7 years ago

I see this error in less than 60 seconds after starting xmr-stak-nvidia, I only see this error so It doesnt work at all. Compiled from git. I made a symlink for cuda, had some problems. Dont know if it matters. My GPU is not overclocked. X kind of hangs when I start it.

lrwxrwxrwx 1 root root 17 Oct 4 15:00 /usr/local/cuda -> /usr/include/cuda/

GPU: GTX980 Dist: Slackware64-current NVIDIA Driver Version: 384.90

Linux bifrost 4.13.4 #1 SMP Sat Sep 30 21:36:54 CEST 2017 x86_64 Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz GenuineIntel GNU/Linux

01:00.0 VGA compatible controller: NVIDIA Corporation GM204 [GeForce GTX 980] (rev a1)

CLANG: clang version 5.0.0 (tags/RELEASE_500/final) Target: x86_64-unknown-linux-gnu Thread model: posix InstalledDir: /usr/bin

__config.txt___

"gpu_threads_conf" : [ { "index" : 0, "threads" : 16, "blocks" : 48, "bfactor" : 0, "bsleep" : 0, "affine_to_cpu" : false}, ],

____output_____

RESULT REPORT Difficulty : 25000 Good results : 0 / 1 (0.0 %) Pool-side hashes : 0

Top 10 best results found: | 0 | 0 | 1 | 0 | | 2 | 0 | 3 | 0 | | 4 | 0 | 5 | 0 | | 6 | 0 | 7 | 0 | | 8 | 0 | 9 | 0 |

____built with:_____ export CC=/usr/bin/clang export CXX=/usr/bin/clang++ cmake . -DCUDA_COMPILER=clang make install

____THE STRUGGLE IS REAL___ --> Deleting empty directory /usr/share/cuda/libnsight/features/com.nvidia.viper.ide.feature_8.0.0.201701130639/META-INF/maven/com.nvidia.viper/com.nvidia.viper.ide.feature/ --> Deleting empty directory /usr/share/cuda/libnsight/features/com.nvidia.viper.ide.feature_8.0.0.201701130639/META-INF/maven/com.nvidia.viper/ --> Deleting empty directory /usr/share/cuda/libnsight/features/com.nvidia.viper.ide.feature_8.0.0.201701130639/META-INF/maven/ --> Deleting empty directory /usr/share/cuda/libnsight/features/com.nvidia.viper.ide.feature_8.0.0.201701130639/META-INF/

export CC=/usr/bin/clang export CXX=/usr/bin/clang++ export PATH=$PATH:/opt/cudatoolkit/bin export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/cudatoolkit/lib cmake . -DCUDA_COMPILER=clang

-- The C compiler identification is Clang 5.0.0 -- The CXX compiler identification is Clang 5.0.0 -- Check for working C compiler: /usr/bin/clang -- Check for working C compiler: /usr/bin/clang -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: /usr/bin/clang++ -- Check for working CXX compiler: /usr/bin/clang++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done -- Looking for pthread.h -- Looking for pthread.h - found -- Looking for pthread_create -- Looking for pthread_create - not found -- Looking for pthread_create in pthreads -- Looking for pthread_create in pthreads - not found -- Looking for pthread_create in pthread -- Looking for pthread_create in pthread - found -- Found Threads: TRUE
-- Found CUDA: /opt/cudatoolkit (found suitable version "9.0", minimum required is "7.5") -- Found OpenSSL: /usr/lib64/libcrypto.so (found version "1.0.2l") -- Configuring done -- Generating done -- Build files have been written to: /opt/xmr-test/xmr-stak-nvidia root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make install Scanning dependencies of target xmr-stak-nvidiaCrypto [ 4%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_blake256.c.o [ 9%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_groestl.c.o [ 14%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_jh.c.o [ 19%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_keccak.c.o [ 23%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_skein.c.o [ 28%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/soft_aes.c.o [ 33%] Linking C static library libxmr-stak-nvidiaCrypto.a [ 33%] Built target xmr-stak-nvidiaCrypto Scanning dependencies of target xmr-stak-nvidiaCuda [ 38%] Building CXX object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o clang-5.0: error: cannot find libdevice for sm_20. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_30. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_35. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_37. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_50. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_52. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_60. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_61. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. clang-5.0: error: cannot find libdevice for sm_62. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. make[2]: [CMakeFiles/xmr-stak-nvidiaCuda.dir/build.make:63: CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o] Error 1 make[1]: [CMakeFiles/Makefile2:105: CMakeFiles/xmr-stak-nvidiaCuda.dir/all] Error 2 make: *** [Makefile:130: all] Error 2 root@bifrost:/opt/xmr-test/xmr-stak-nvidia# ls -ltr /usr/local/cuda lrwxrwxrwx 1 root root 16 Oct 5 23:28 /usr/local/cuda -> /opt/cudatoolkit/ root@bifrost:/opt/xmr-test/xmr-stak-nvidia# cmake . -DCUDA_COMPILER=clang -nocudalib CMake Error: The source directory "/opt/xmr-test/xmr-stak-nvidia/-nocudalib" does not exist. Specify --help for usage, or press the help button on the CMake GUI. root@bifrost:/opt/xmr-test/xmr-stak-nvidia# cmake . -DCUDA_COMPILER=nvcc -- Configuring done -- Generating done -- Build files have been written to: /opt/xmr-test/xmr-stak-nvidia root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make install Scanning dependencies of target xmr-stak-nvidiaCrypto [ 4%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_blake256.c.o [ 9%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_groestl.c.o [ 14%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_jh.c.o [ 19%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_keccak.c.o [ 23%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_skein.c.o [ 28%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/soft_aes.c.o [ 33%] Linking C static library libxmr-stak-nvidiaCrypto.a [ 33%] Built target xmr-stak-nvidiaCrypto [ 38%] Building NVCC (Device) object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o nvcc fatal : Unsupported gpu architecture 'compute_20' CMake Error at xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o.Release.cmake:222 (message): Error generating /opt/xmr-test/xmr-stak-nvidia/CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/./xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o

make[2]: [CMakeFiles/xmr-stak-nvidiaCuda.dir/build.make:72: CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o] Error 1 make[1]: [CMakeFiles/Makefile2:104: CMakeFiles/xmr-stak-nvidiaCuda.dir/all] Error 2 make: *** [Makefile:130: all] Error 2 root@bifrost:/opt/xmr-test/xmr-stak-nvidia# ls bin/ root@bifrost:/opt/xmr-test/xmr-stak-nvidia# cmake . -DCUDA_COMPILER=nvcc -DCUDA_ARCH=52 -- Configuring done -- Generating done -- Build files have been written to: /opt/xmr-test/xmr-stak-nvidia root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make install [ 33%] Built target xmr-stak-nvidiaCrypto [ 38%] Building NVCC (Device) object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o nvcc fatal : The version ('50000') of the host compiler ('clang') is not supported CMake Error at xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o.Release.cmake:222 (message): Error generating /opt/xmr-test/xmr-stak-nvidia/CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/./xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o

make[2]: [CMakeFiles/xmr-stak-nvidiaCuda.dir/build.make:72: CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/xmr-stak-nvidiaCuda_generated_cuda_extra.cu.o] Error 1 make[1]: [CMakeFiles/Makefile2:104: CMakeFiles/xmr-stak-nvidiaCuda.dir/all] Error 2 make: [Makefile:130: all] Error 2 root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make clean root@bifrost:/opt/xmr-test/xmr-stak-nvidia# cmake . -DCUDA_COMPILER=clang -- Configuring done -- Generating done -- Build files have been written to: /opt/xmr-test/xmr-stak-nvidia root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make install Scanning dependencies of target xmr-stak-nvidiaCrypto [ 4%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_blake256.c.o [ 9%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_groestl.c.o [ 14%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_jh.c.o [ 19%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_keccak.c.o [ 23%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_skein.c.o [ 28%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/soft_aes.c.o [ 33%] Linking C static library libxmr-stak-nvidiaCrypto.a [ 33%] Built target xmr-stak-nvidiaCrypto Scanning dependencies of target xmr-stak-nvidiaCuda [ 38%] Building CXX object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o clang-5.0: error: cannot find libdevice for sm_52. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. make[2]: [CMakeFiles/xmr-stak-nvidiaCuda.dir/build.make:63: CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o] Error 1 make[1]: [CMakeFiles/Makefile2:105: CMakeFiles/xmr-stak-nvidiaCuda.dir/all] Error 2 make: [Makefile:130: all] Error 2 root@bifrost:/opt/xmr-test/xmr-stak-nvidia# cmake . -DCUDA_COMPILER=clang -DCUDA_ARCH=62 -- Configuring done -- Generating done -- Build files have been written to: /opt/xmr-test/xmr-stak-nvidia root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make install [ 33%] Built target xmr-stak-nvidiaCrypto [ 38%] Building CXX object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o clang-5.0: error: cannot find libdevice for sm_62. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. make[2]: [CMakeFiles/xmr-stak-nvidiaCuda.dir/build.make:63: CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o] Error 1 make[1]: [CMakeFiles/Makefile2:105: CMakeFiles/xmr-stak-nvidiaCuda.dir/all] Error 2 make: [Makefile:130: all] Error 2 root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make clean root@bifrost:/opt/xmr-test/xmr-stak-nvidia# cmake . -DCUDA_COMPILER=clang -DCUDA_ARCH=61 -- Configuring done -- Generating done -- Build files have been written to: /opt/xmr-test/xmr-stak-nvidia root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make install [ 4%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_blake256.c.o [ 9%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_groestl.c.o [ 14%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_jh.c.o [ 19%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_keccak.c.o [ 23%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/c_skein.c.o [ 28%] Building C object CMakeFiles/xmr-stak-nvidiaCrypto.dir/crypto/soft_aes.c.o [ 33%] Linking C static library libxmr-stak-nvidiaCrypto.a [ 33%] Built target xmr-stak-nvidiaCrypto [ 38%] Building CXX object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o clang-5.0: error: cannot find libdevice for sm_61. Provide path to different CUDA installation via --cuda-path, or pass -nocudalib to build without linking with libdevice. make[2]: [CMakeFiles/xmr-stak-nvidiaCuda.dir/build.make:63: CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o] Error 1 make[1]: [CMakeFiles/Makefile2:105: CMakeFiles/xmr-stak-nvidiaCuda.dir/all] Error 2 make: [Makefile:130: all] Error 2

Yeah I just downloaded Fedora runfile for cudatoolkit and installer in /usr with --override, this is a great idea :D

cmake . -DCUDA_COMPILER=clang root@bifrost:/opt/xmr-test/xmr-stak-nvidia# make install [ 33%] Built target xmr-stak-nvidiaCrypto [ 38%] Building CXX object CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o c++: error: unrecognized command line option ‘--cuda-path=/usr’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_20’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_30’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_35’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_37’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_50’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_52’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_60’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_61’ c++: error: unrecognized command line option ‘--cuda-gpu-arch=sm_62’ make[2]: [CMakeFiles/xmr-stak-nvidiaCuda.dir/build.make:63: CMakeFiles/xmr-stak-nvidiaCuda.dir/nvcc_code/cuda_core.cu.o] Error 1 make[1]: [CMakeFiles/Makefile2:105: CMakeFiles/xmr-stak-nvidiaCuda.dir/all] Error 2 make: *** [Makefile:130: all] Error 2 root@bifrost:/opt/xmr-test/xmr-stak-nvidia#

seems like cuda 9.0 is missing sm_52 and stuff in /usr/include/cuda?

git checkout 3326455 cmake . -DCUDA_COMPILER=clang make install

RESULT REPORT Difficulty : 3473 Good results : 0 / 33 (0.0 %) Pool-side hashes : 0

Top 10 best results found: | 0 | 0 | 1 | 0 | | 2 | 0 | 3 | 0 | | 4 | 0 | 5 | 0 | | 6 | 0 | 7 | 0 | | 8 | 0 | 9 | 0 |

hmmm

ghost commented 7 years ago

FRESH UBUNTU 17.04 INSTALL

nvcc -V

7  apt-get update
8  apt-get upgrade
9  apt-get dist-upgrade

10 apt-get install mumble 11 apt-get -y install gcc g++ build-essential automake linux-headers-$(uname -r) git gawk libcurl4-openssl-dev libjansson-dev xorg libc++-dev libgmp-dev python-dev 24 apt-get install nvidia-cuda-dev nvidia-cuda-toolkit libmicrohttpd-dev libssl-dev cmake cmake-curses-gui build-essential 65 git clone https://github.com/fireice-uk/xmr-stak-nvidia.git 66 cd xmr-stak-nvidia/ 50 export CC=/usr/bin/clang-3.8 51 export CXX=/usr/bin/clang++-3.8 cmake . 49 make install

cd bin/ vi config.txt ./xmr-stak-nvidia

XMR-Stak-NVIDIA mining software, NVIDIA Version. NVIDIA mining code was written by KlausT and psychocrypt. Brought to you by fireice_uk under GPLv3.

Configurable dev donation level is set to 1.0 %

You can use following keys to display reports: 'h' - hashrate 'r' - results 'c' - connection

[2017-10-09 19:17:01] : Connecting to pool pool.supportxmr.com:7777 ... No CUDA device found! [2017-10-09 19:17:01] : Setup failed for GPU 0. Exitting.

(it screws up the terminal so I have to write "reset")

DONT DO ANYTHING BELOW, PLEASE SEE MY LAST POST DONT DO ANYTHING BELOW, PLEASE SEE MY LAST POST DONT DO ANYTHING BELOW, PLEASE SEE MY LAST POST

remove all nvidia packages ,skip this if your system is fresh installed
sudo apt-get remove nvidia* && sudo apt autoremove
install some packages for build kernel:
sudo apt-get install dkms build-essential linux-headers-generic
now block and disable nouveau kernel driver:
sudo vim /etc/modprobe.d/blacklist.conf

Insert follow lines to the blacklist.conf: blacklist nouveau blacklist lbm-nouveau options nouveau modeset=0 alias nouveau off alias lbm-nouveau off

save and exit. Disable the Kernel nouveau by typing the following commands(nouveau-kms.conf may not exist,it is ok): echo options nouveau modeset=0 | sudo tee -a /etc/modprobe.d/nouveau-kms.conf build the new kernel by: sudo update-initramfs -u reboot

nvidia.com, download latest driver .run file chmod+x ./NV*.run

nvidia.com cuda toolkit, download .deb for ubuntu17.04 just double click it

+-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 1140 G /usr/lib/xorg/Xorg 238MiB | | 0 1720 G /usr/bin/compiz 58MiB | +-----------------------------------------------------------------------------+ root@bifrost:/opt/xmr-stak-nvidia# nvcc -V

ok so this is great, its working now. root@bifrost:/opt/xmr-stak-nvidia# nvcc -V The program 'nvcc' is currently not installed. You can install it by typing: apt install nvidia-cuda-toolkit

this is bad since I installed nvidia cuda

psychocrypt commented 7 years ago

Could you please link which version is not effected? Do you mean the release or a commit within the master or sev branch.

ghost commented 7 years ago

I dont know, I reinstalled my system. Right now this is just the latest master.

Still cannot run xmr-stack-nvidia properly. Dont know if its my config: { "index" : 0, "threads" : 16, "blocks" : 32, "bfactor" : 8, "bsleep" : 25, "affine_to_cpu" : false},

Tried some other but just getting error above.

root@bifrost:/opt# apt-get install clinfo root@bifrost:/opt/xmr-stak-nvidia/bin# clinfo Number of platforms 0

:(

psychocrypt commented 7 years ago

Sry I missed that so many user writing into this thread. @frankcoderL please link the last version which can be used without issues.

ghost commented 7 years ago

I finally got it to work.

[2017-10-09 23:43:02] : Result accepted by the pool.

7  apt-get update
8  apt-get upgrade
9  apt-get dist-upgrade

11 apt-get -y install gcc g++ build-essential automake linux-headers-$(uname -r) git gawk libcurl4-openssl-dev libjansson-dev xorg libc++-dev libgmp-dev python-dev 24 apt-get install nvidia-cuda-dev nvidia-cuda-toolkit libmicrohttpd-dev libssl-dev cmake cmake-curses-gui build-essential 65 git clone https://github.com/fireice-uk/xmr-stak-nvidia.git 66 cd xmr-stak-nvidia/ 50 export CC=/usr/bin/clang-3.8 51 export CXX=/usr/bin/clang++-3.8 cmake . 49 make install

cd bin/ vi config.txt ./xmr-stak-nvidia

apt remove --purge nvidia apt install nvidia-common apt install nvidia-current apt remove --purge xserver-xorg-video-nouveau reboot and let aptitude disable nouveau DONT DISABLE NOUVEAU MANUALLY* (do you want to know how much time ive wasted?)

download latest nvidia.run fil and install it by systemctl isolate multi-user.target ctrl + f1 and login chmod +x NV.run ./NV.run

hopefully nvidia will prompt to install dkms things reboot and run xmr-stak-nvidia

somehow that worked for me.

for the love of god dont install nvidia-cuda-toolkit after this, you will bork things upp prolly leave everything with cuda and nvidia in aptitude for now D:

frankcoderL commented 7 years ago

@ psychocrypt The problem that trite2k3 describes (driver problem that gives a GPU error on xmr-stak-nvidia) is not the same as mine.

On my system:

The git clone + compilation of 28/07/2017 does work fine (still up and running)
The current git clone + compilation gives a GPU error all the time I don't know what causes this issue and what code commit (if this is a cause) causes this GPU error yet.

ghost commented 7 years ago

What are you talking about? We had the same problem "gpu compute error", I just solved it. Probably with clang 3.8 instead of 5.0.

frankcoderL commented 7 years ago

Using clang 3.8 instead of clang 5.0 is a workaround and not a solution. But it is an importent clue in what is wrong to find a solution for the "gpu compute error".

Since i'm using Arch Linux that uses a rolling-release system I prefer to use the latest stable release, and that is clang 5.0 at this moment.

Because we use different Linux distributions I might have an other problem than you. I thought first it was a nvidia-driver problem but it might also be a clang compiling problem as you pointed out, the future will tell.

psychocrypt commented 7 years ago

@frankcoderL Could you write me a PM on reddit: my name there is psychocrypt We need to check if the next release solved the issu.

trite2k3 commented 7 years ago

I have now successfully built it on two different machine using slackware64-current with latest stable kernel from kernel.org and NVIDIA drivers from .run file.

Dont know if it was because clang 3.8 or if psycocrypt solved it in latest release.

Heres what I did:

git clone https://github.com/fireice-uk/xmr-stak-nvidia.git

install microhttpd from slackbuilds

install clang 3-8 from slackware 14.2

ln -s /usr/bin/clang-3.8 /usr/bin/clang

export CC=/usr/bin/clang

export CXX=/usr/bin/clang++

export PATH=/usr/local/sbin:/usr/sbin:/sbin:/usr/local/bin:/usr/bin:/bin:/usr/games:/usr/share/cuda/bin:/usr/lib64/qt/bin:/usr/share/texmf/bin:/usr/share/cuda/:/usr/include/cuda:/usr/share/cuda/bin:/usr/share/cuda/lib64

export LD_LIBRARY_PATH=/usr/share/cuda/:/usr/include/cuda:/usr/share/cuda/bin:/usr/share/cuda/lib64

cmake . -DCUDA_SDK_ROOT_DIR=/usr/include/cuda -DCUDA_TOOLKIT_ROOT_DIR=/usr/share/cuda

vi /usr/include/c++/7/bits/std_abs.h

change line 99 from:

if !defined(__STRICT_ANSI__) && defined(_GLIBCXX_USE_FLOAT128)

to:

if defined(_GLIBCXX_USE_FLOAT128) && !defined(__STRICT_ANSI) && !defined(CUDACC__)

make install

frankcoderL commented 7 years ago

@psychocrypt, I sended a PM on reddit, my name is frankcoderL on reddit. Sorry for the delay.

trite2k3 commented 7 years ago

this https://people.gnome.org/~ole/includeos/ubuntu-17.10-c++-20171029.patch

fireice-uk / xmr-stak-nvidia

Error GPU Compute Error #124

[GPU COMPUTE ERROR]

You can use following keys to display reports: 'h' - hashrate 'r' - results 'c' - connection

if !defined(__STRICT_ANSI__) && defined(_GLIBCXX_USE_FLOAT128)

if defined(_GLIBCXX_USE_FLOAT128) && !defined(__STRICT_ANSI) && !defined(CUDACC__)