cbuchner1 / CudaMiner

a CUDA accelerated litecoin mining application based on pooler's CPU miner
Other
692 stars 304 forks source link

Ubuntu 14.04 Pointer Issue #112

Closed webbrandon closed 10 years ago

webbrandon commented 10 years ago

Hello,

I am compiling for a Ubuntu 14.04 x64 system with the 331.38 NVidia Driver for a GeForce GTX 670 card. I am using the new driver support for NVIDIA (System Settings > Software Settings > Additional Drivers) and the new sudo apt-get install nvidia-cuda-toolkit availability. The cuda package installs to /usr/lib/nvidia-cuda-toolkit instead of the normal /usr/local/cuda.

How can I point this for CudaMiner to compile correctly? Currently I get this: /usr/local/cuda/bin/nvcc -O3 -Xptxas "-abi=no -v" -arch=compute_10 --maxrregcount=64 --ptxas-options=-v -I./compat/jansson -o sha256.o -c sha256.cu /bin/bash: /usr/local/cuda/bin/nvcc: No such file or directory make[2]: *** [salsa_kernel.o] Error 127

SteveGotthardt commented 10 years ago

Try this in directory with the cudaminer source: make clean ./autogen.sh ./configure --with-cuda=/usr/lib/nvidia-cuda/toolkit make (and optionally: sudo make install)

On Tue, Mar 11, 2014 at 3:28 PM, webbrandon notifications@github.comwrote:

Hello,

I am compiling for a Ubuntu 14.04 x64 system with the 331.48 NVidia Driver for a GeForce GTX 670 card. I am using the new driver support for NVIDIA (System Settings > Software Settings > Additional Drivers) and the new sudo apt-get install nvidia-cuda-toolkit availability. The cuda package installs to /usr/lib/nvidia-cuda-toolkit instead of the normal /usr/local/cuda.

How can I point this for CudaMiner to compile correctly? Currently I get this: /usr/local/cuda/bin/nvcc -O3 -Xptxas "-abi=no -v" -arch=compute_10 --maxrregcount=64 --ptxas-options=-v -I./compat/jansson -o sha256.o -c sha256.cu /bin/bash: /usr/local/cuda/bin/nvcc: No such file or directory make[2]: *\ [salsa_kernel.o] Error 127

Reply to this email directly or view it on GitHubhttps://github.com/cbuchner1/CudaMiner/issues/112 .

SteveGotthardt commented 10 years ago

Sorry, dash instead of slash (or point it to where cuda is on your install) ./configure --with-cuda=/usr/lib/nvidia-cuda-toolkit

On Tue, Mar 11, 2014 at 3:33 PM, Steve Gotthardt gotthardt@gmail.comwrote:

Try this in directory with the cudaminer source: make clean ./autogen.sh ./configure --with-cuda=/usr/lib/nvidia-cuda/toolkit make (and optionally: sudo make install)

On Tue, Mar 11, 2014 at 3:28 PM, webbrandon notifications@github.comwrote:

Hello,

I am compiling for a Ubuntu 14.04 x64 system with the 331.48 NVidia Driver for a GeForce GTX 670 card. I am using the new driver support for NVIDIA (System Settings > Software Settings > Additional Drivers) and the new sudo apt-get install nvidia-cuda-toolkit availability. The cuda package installs to /usr/lib/nvidia-cuda-toolkit instead of the normal /usr/local/cuda.

How can I point this for CudaMiner to compile correctly? Currently I get this: /usr/local/cuda/bin/nvcc -O3 -Xptxas "-abi=no -v" -arch=compute_10 --maxrregcount=64 --ptxas-options=-v -I./compat/jansson -o sha256.o -c sha256.cu /bin/bash: /usr/local/cuda/bin/nvcc: No such file or directory make[2]: *\ [salsa_kernel.o] Error 127

Reply to this email directly or view it on GitHubhttps://github.com/cbuchner1/CudaMiner/issues/112 .

webbrandon commented 10 years ago

It compiled, thank you. I will write a tutorial for 14.04 on askubuntu.com and reference here later.

webbrandon commented 10 years ago

BTW not a optimal setup I am only getting 170 khash/s average compared to build made from sources of 280 khash/s average.