"CUDA runtime version 0.0" when running mfaktc.exe

froderik · 2016-10-12, 12:10

Dear all. Hello.
I have an issue with running mfaktc.exe on my Ubuntu 16.04.1 LTS.
I have successfully compiled mfaktc-0.21 from sources. But when I run mfaktc.exe I receive the next error (please see output):

mfaktc v0.21 (64bit built)

Compiletime options
THREADS_PER_BLOCK 256
SIEVE_SIZE_LIMIT 32kiB
SIEVE_SIZE 193154bits
SIEVE_SPLIT 250
MORE_CLASSES enabled

Runtime options
SievePrimes 25000
SievePrimesAdjust 1
SievePrimesMin 5000
SievePrimesMax 100000
NumStreams 3
CPUStreams 3
GridSize 3
GPU Sieving enabled
GPUSievePrimes 82486
GPUSieveSize 64Mi bits
GPUSieveProcessSize 16Ki bits
Checkpoints enabled
CheckpointDelay 30s
WorkFileAddDelay 600s
Stages enabled
StopAfterFactor bitlevel
PrintMode full
V5UserID (none)
ComputerID (none)
AllowSleep no
TimeStampInResults no

CUDA version info
binary compiled for CUDA 8.0
CUDA runtime version 0.0
CUDA driver version 8.0
ERROR: CUDA runtime version must match the CUDA toolkit version used during compile!

Also command "ldd mfaktc.exe" shows me :

linux-vdso.so.1 => (0x00007fffab3af000)
libcudart.so.8.0 => /usr/local/cuda-8.0/lib64/libcudart.so.8.0 (0x00007f6f2a503000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f6f2a1d1000)
libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f6f29e4e000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f6f29c38000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f6f2986f000)
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f6f2966a000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f6f2944d000)
librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f6f29245000)
/lib64/ld-linux-x86-64.so.2 (0x00005574ccbe6000)

"env | grep cuda" gives me :

LD_LIBRARY_PATH=:/usr/local/cuda-8.0/lib64
PATH=/home/falifirenko/perl5/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/node/bin:/opt/idea-IC-145.1617.8/bin:/opt/pycharm-community-2016.1.2/bin:/usr/local/cuda-8.0/bin

"nvidia-smi" command gives me:

Wed Oct 12 15:08:02 2016
+------------------------------------------------------+
| NVIDIA-SMI 361.42 Driver Version: 361.42 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GT 650M Off | 0000:01:00.0 N/A | N/A |
| N/A 51C P8 N/A / N/A | 349MiB / 2047MiB | N/A Default |
+-------------------------------+----------------------+----------------------+

Please help me to solve this problem.

Thanks in advance !

Mark Rose · 2016-10-12, 15:29

If you run `sudo nvidia-modprobe` does it fix it?

chris2be8 · 2016-10-12, 16:08

What does lspci show? In particular check the nvidia kernel driver is in use as follows:

Code:

4core:~ # lspci -v | grep -i nvidia
01:00.0 VGA compatible controller: NVIDIA Corporation Device 13c2 (rev a1) (prog-if 00 [VGA controller])
        Kernel driver in use: nvidia
        Kernel modules: nouveau, nvidia
01:00.1 Audio device: NVIDIA Corporation Device 0fbb (rev a1)

To make my system use the nvidia driver I had to blacklist the nouveau driver by adding it to /etc/modprobe.d/blacklist.conf (if it's not already there) and then run "update-initramfs -u" (the system ignores /etc/modprobe.d/blacklist.conf until you do that). After a reboot lspci -v showed the card was using the nvidia driver.

Chris

bgbeuning · 2016-10-25, 02:17

I am having the same problem. How did you fix it?

The machine has both cuda7.5 and cuda8.0 installed.
And both gcc-4.9 and gcc-5.4 .

I use gcc-4.9 with nvcc when compiling mfaktc.

nvidia-modprobe did not help.

nvidia-smi gives an error when running. See below.

Here are details

mfaktc v0.21 (64bit built)

Compiletime options
THREADS_PER_BLOCK 256
SIEVE_SIZE_LIMIT 32kiB
SIEVE_SIZE 193154bits
SIEVE_SPLIT 250
MORE_CLASSES enabled

Runtime options
SievePrimes 25000
SievePrimesAdjust 1
SievePrimesMin 5000
SievePrimesMax 100000
NumStreams 3
CPUStreams 3
GridSize 3
GPU Sieving enabled
GPUSievePrimes 82486
GPUSieveSize 64Mi bits
GPUSieveProcessSize 16Ki bits
Checkpoints enabled
CheckpointDelay 30s
WorkFileAddDelay 600s
Stages enabled
StopAfterFactor bitlevel
PrintMode full
V5UserID (none)
ComputerID (none)
AllowSleep no
TimeStampInResults no

CUDA version info
binary compiled for CUDA 8.0
CUDA runtime version 0.0
CUDA driver version 8.0
ERROR: CUDA runtime version must match the CUDA toolkit version used during compile!

Instead of setting LD_LIBRARY_PATH, I use ldconfig

bgb@i5-6500:~/gpu/0$ ls /etc/ld.so.conf.d
amdapp_x86_64.conf fakeroot-x86_64-linux-gnu.conf x86_64-linux-gnu.conf
amdapp_x86.conf i386-linux-gnu.conf x86_64-linux-gnu_EGL.conf
amdgpu.conf i386-linux-gnu_EGL.conf x86_64-linux-gnu_GL.conf
cuda75.conf i386-linux-gnu_GL.conf zz_i386-biarch-compat.conf
cuda80.conf libc.conf
bgb@i5-6500:~/gpu/0$ cat /etc/ld.so.conf.d/cuda*
/usr/local/cuda-7.5/targets/x86_64-linux/lib
/usr/local/cuda-8.0/lib64
bgb@i5-6500:~/gpu/0$

bgb@i5-6500:~/gpu/0$ ldd ../mfaktc-0.21/mfaktc.exe
linux-vdso.so.1 => (0x00007ffea6fbc000)
libcudart.so.8.0 => /usr/local/cuda-8.0/lib64/libcudart.so.8.0 (0x00007f30dffa5000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f30dfc9c000)
libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f30df919000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f30df703000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f30df33a000)
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f30df135000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f30def18000)
librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f30ded10000)
/lib64/ld-linux-x86-64.so.2 (0x000055650a453000)

bgb@i5-6500:~/gpu/0$ nvidia-smi
Failed to initialize NVML: Driver/library version mismatch
bgb@i5-6500:~/gpu/0$

bgb@i5-6500:~/gpu/0$ lspci -v | grep -i nvidia
01:00.0 VGA compatible controller: NVIDIA Corporation GM204 [GeForce GTX 970] (rev a1) (prog-if 00 [VGA controller])
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_361, nvidia_drm, nvidia
01:00.1 Audio device: NVIDIA Corporation GM204 High Definition Audio Controller (rev a1)
07:00.0 VGA compatible controller: NVIDIA Corporation GM204 [GeForce GTX 970] (rev a1) (prog-if 00 [VGA controller])
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_361, nvidia_drm, nvidia
07:00.1 Audio device: NVIDIA Corporation GM204 High Definition Audio Controller (rev a1)

bgbeuning · 2016-10-30, 15:29

I had a version 361 driver installed and it wanted the 367 version.
dmesg(1) had lots of messages giving both version numbers.
The nvidia "run" script did not do a complete job in installing the driver.
After running the 367 run script, it was still using the 361 driver.
The apt-get install path did install 367 right so it would run.

2016-10-12, 12:10	#1
froderik Oct 2016 1₂ Posts	"CUDA runtime version 0.0" when running mfaktc.exe Dear all. Hello. I have an issue with running mfaktc.exe on my Ubuntu 16.04.1 LTS. I have successfully compiled mfaktc-0.21 from sources. But when I run mfaktc.exe I receive the next error (please see output): mfaktc v0.21 (64bit built) Compiletime options THREADS_PER_BLOCK 256 SIEVE_SIZE_LIMIT 32kiB SIEVE_SIZE 193154bits SIEVE_SPLIT 250 MORE_CLASSES enabled Runtime options SievePrimes 25000 SievePrimesAdjust 1 SievePrimesMin 5000 SievePrimesMax 100000 NumStreams 3 CPUStreams 3 GridSize 3 GPU Sieving enabled GPUSievePrimes 82486 GPUSieveSize 64Mi bits GPUSieveProcessSize 16Ki bits Checkpoints enabled CheckpointDelay 30s WorkFileAddDelay 600s Stages enabled StopAfterFactor bitlevel PrintMode full V5UserID (none) ComputerID (none) AllowSleep no TimeStampInResults no CUDA version info binary compiled for CUDA 8.0 CUDA runtime version 0.0 CUDA driver version 8.0 ERROR: CUDA runtime version must match the CUDA toolkit version used during compile! Also command "ldd mfaktc.exe" shows me : linux-vdso.so.1 => (0x00007fffab3af000) libcudart.so.8.0 => /usr/local/cuda-8.0/lib64/libcudart.so.8.0 (0x00007f6f2a503000) libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f6f2a1d1000) libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f6f29e4e000) libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f6f29c38000) libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f6f2986f000) libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f6f2966a000) libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f6f2944d000) librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f6f29245000) /lib64/ld-linux-x86-64.so.2 (0x00005574ccbe6000) "env \| grep cuda" gives me : LD_LIBRARY_PATH=:/usr/local/cuda-8.0/lib64 PATH=/home/falifirenko/perl5/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/node/bin:/opt/idea-IC-145.1617.8/bin:/opt/pycharm-community-2016.1.2/bin:/usr/local/cuda-8.0/bin "nvidia-smi" command gives me: Wed Oct 12 15:08:02 2016 +------------------------------------------------------+ \| NVIDIA-SMI 361.42 Driver Version: 361.42 \| \|-------------------------------+----------------------+----------------------+ \| GPU Name Persistence-M\| Bus-Id Disp.A \| Volatile Uncorr. ECC \| \| Fan Temp Perf Pwr:Usage/Cap\| Memory-Usage \| GPU-Util Compute M. \| \|===============================+======================+======================\| \| 0 GeForce GT 650M Off \| 0000:01:00.0 N/A \| N/A \| \| N/A 51C P8 N/A / N/A \| 349MiB / 2047MiB \| N/A Default \| +-------------------------------+----------------------+----------------------+ Please help me to solve this problem. Thanks in advance !

2016-10-12, 16:08	#3
chris2be8 Sep 2009 2,333 Posts	What does lspci show? In particular check the nvidia kernel driver is in use as follows: Code: 4core:~ # lspci -v \| grep -i nvidia 01:00.0 VGA compatible controller: NVIDIA Corporation Device 13c2 (rev a1) (prog-if 00 [VGA controller]) Kernel driver in use: nvidia Kernel modules: nouveau, nvidia 01:00.1 Audio device: NVIDIA Corporation Device 0fbb (rev a1) To make my system use the nvidia driver I had to blacklist the nouveau driver by adding it to /etc/modprobe.d/blacklist.conf (if it's not already there) and then run "update-initramfs -u" (the system ignores /etc/modprobe.d/blacklist.conf until you do that). After a reboot lspci -v showed the card was using the nvidia driver. Chris

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Patch] "Test/Primenet" prompts improvements on console version	Explorer09	Software	2	2017-03-09 04:14
"LateX" version of factorX?	kracker	Operation Billion Digits	8	2012-12-04 18:35
Program not running: icon "blink"	Rassilon6	Software	3	2007-02-02 01:24
Prime95 Version 24.13 "Feature"	RMAC9.5	Software	2	2006-03-24 21:12
Running the "Difference Method" on Mathematica	jinydu	Programming	6	2005-10-01 00:09

2016-10-12, 15:29	#2
Mark Rose "/X\(‘-‘)/X\" Jan 2013 2³×3²×41 Posts	If you run `sudo nvidia-modprobe` does it fix it?

2016-10-25, 02:17	#4
bgbeuning Dec 2014 3×5×17 Posts	I am having the same problem. How did you fix it? The machine has both cuda7.5 and cuda8.0 installed. And both gcc-4.9 and gcc-5.4 . I use gcc-4.9 with nvcc when compiling mfaktc. nvidia-modprobe did not help. nvidia-smi gives an error when running. See below. Here are details mfaktc v0.21 (64bit built) Compiletime options THREADS_PER_BLOCK 256 SIEVE_SIZE_LIMIT 32kiB SIEVE_SIZE 193154bits SIEVE_SPLIT 250 MORE_CLASSES enabled Runtime options SievePrimes 25000 SievePrimesAdjust 1 SievePrimesMin 5000 SievePrimesMax 100000 NumStreams 3 CPUStreams 3 GridSize 3 GPU Sieving enabled GPUSievePrimes 82486 GPUSieveSize 64Mi bits GPUSieveProcessSize 16Ki bits Checkpoints enabled CheckpointDelay 30s WorkFileAddDelay 600s Stages enabled StopAfterFactor bitlevel PrintMode full V5UserID (none) ComputerID (none) AllowSleep no TimeStampInResults no CUDA version info binary compiled for CUDA 8.0 CUDA runtime version 0.0 CUDA driver version 8.0 ERROR: CUDA runtime version must match the CUDA toolkit version used during compile! Instead of setting LD_LIBRARY_PATH, I use ldconfig bgb@i5-6500:~/gpu/0$ ls /etc/ld.so.conf.d amdapp_x86_64.conf fakeroot-x86_64-linux-gnu.conf x86_64-linux-gnu.conf amdapp_x86.conf i386-linux-gnu.conf x86_64-linux-gnu_EGL.conf amdgpu.conf i386-linux-gnu_EGL.conf x86_64-linux-gnu_GL.conf cuda75.conf i386-linux-gnu_GL.conf zz_i386-biarch-compat.conf cuda80.conf libc.conf bgb@i5-6500:~/gpu/0$ cat /etc/ld.so.conf.d/cuda* /usr/local/cuda-7.5/targets/x86_64-linux/lib /usr/local/cuda-8.0/lib64 bgb@i5-6500:~/gpu/0$ bgb@i5-6500:~/gpu/0$ ldd ../mfaktc-0.21/mfaktc.exe linux-vdso.so.1 => (0x00007ffea6fbc000) libcudart.so.8.0 => /usr/local/cuda-8.0/lib64/libcudart.so.8.0 (0x00007f30dffa5000) libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f30dfc9c000) libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f30df919000) libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f30df703000) libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f30df33a000) libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f30df135000) libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f30def18000) librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f30ded10000) /lib64/ld-linux-x86-64.so.2 (0x000055650a453000) bgb@i5-6500:~/gpu/0$ nvidia-smi Failed to initialize NVML: Driver/library version mismatch bgb@i5-6500:~/gpu/0$ bgb@i5-6500:~/gpu/0$ lspci -v \| grep -i nvidia 01:00.0 VGA compatible controller: NVIDIA Corporation GM204 [GeForce GTX 970] (rev a1) (prog-if 00 [VGA controller]) Kernel driver in use: nvidia Kernel modules: nvidiafb, nouveau, nvidia_361, nvidia_drm, nvidia 01:00.1 Audio device: NVIDIA Corporation GM204 High Definition Audio Controller (rev a1) 07:00.0 VGA compatible controller: NVIDIA Corporation GM204 [GeForce GTX 970] (rev a1) (prog-if 00 [VGA controller]) Kernel driver in use: nvidia Kernel modules: nvidiafb, nouveau, nvidia_361, nvidia_drm, nvidia 07:00.1 Audio device: NVIDIA Corporation GM204 High Definition Audio Controller (rev a1)

2016-10-30, 15:29	#5
bgbeuning Dec 2014 3·5·17 Posts	I had a version 361 driver installed and it wanted the 367 version. dmesg(1) had lots of messages giving both version numbers. The nvidia "run" script did not do a complete job in installing the driver. After running the 367 run script, it was still using the 361 driver. The apt-get install path did install 367 right so it would run.