I am buying ATI Stream 9270 and Nvidia Tesla C1060, and will try to compare their performance.
Can I install them into the same linux box and run AIT Stream and CUDA programs on it?
Not at the same time, I guess.
Thanks for your reply.
I guess that OpenGL stuff might be a key.
It seems that I would have to make system dual boot for both GPU.
It may be possible to load the proper driver using a dual boot setup, but you may still encounter hardware conflicts. I doubt two main gfx adapters can coexist on a single motherboard unless one is the SLI/Crossfire master of the other.
I already have a dual boot linux box which has a Radeon 4550 and a Geforce 9880 on the same mother boad.
I can run ATI Stream and CUDA programs on each linux system.
I read somewhere that if we have multiple GPUs (all of them must be the same GPU card) and want to run GPGPU program on them, we have to disable SLI/Crossfire or only on GPU can be used.
This is a very interesting thing - that you could do AMD Streams n CUDA at the same time.
So, Does linux support loading multiple display drivers at the same time? (Windows does not seem to like. May b, I am wrong).
What about your X configuration et al?
Thanks,
Sarnath
> So, Does linux support loading multiple display drivers at the same time? (Windows does not seem to like. May b, I am wrong).
It can be easily done in Windows Seven (at least build 7077), but not in XP or Vista.
Hi there,
I've been able to run cal and cuda at the same time, so far with success and proving to be more convenient than my previous dual-boot arrangement, as follows. This is with a Scientific Linux 5.1 64 bit (clone of RHEL) with a Radeon HD 3870 and an Nvidia GTX260. The procedure seemed a bit "hairy" but appears to work (I don't know if you can just install one driver after another in case they interfere with each other about X/opengl stuff). Note I am not a linux expert so consider this "informational" only! The key thing to note is that CAL needs X running with the ATI driver whereas cuda only needs the Nvidia kernel module in place, and indeed Nvidia provide a script to enable one to run cuda programs at runlevel 3. So one does a "minimal" Nvidia install followed by a "regular" ATI install. Note with this setup no cuda-graphics interop is possible.
So, from memory, something like this should work:
Uninstall any existing Nvidia/ATI drivers
Put the ATI card in your primary PCIe slot with a monitor attached
Install linux if you haven't done so already.
Put the Nvidia card(s) in your secondary PCIe slot without a monitor connected
In runlevel 3 as root,
Download the nvidia driver, NVIDIA*.run (version 180.60 in my case)
Extract the contents with ./NVIDIA*.run -x
Go into the usr/src/nv subdirectory of the directory just created
Make and install the nvidia kernel module with "make module" and then "make install"
Now go to usr/lib and copy libcuda.so.180.60 to /lib64
Make links from libcuda.so to libcuda.so.1 and from libcuda.so.1 to libcuda.so.180.60 in /lib64
Copy usr/include/cuda/cuda.h from the directory to /usr/include/cuda/cuda.h
There might be an alternative to some of the above steps (say by running nvidia-installer in advanced mode, dumping everything except the kernel module into /tmp and copying across the library and headers from there).
Next you need to put the script Nvidia provide in their release notes into a suitable place. I'm a total amateur here, but I copied /etc/rc.d/rc.local to /etc/rc.d/rc.mylocal, merged the nvidia material into it, then changed the "S99local" symbolic link in /etc/rc.d/rc5.d to point to rc.mylocal. I may have had to change some of the commands to use the full paths to them also. (If an expert can tell me a better thing to do here I'd be very grateful.)
Then, in runlevel 5, run the ati-driver-installer with the default settings.
Log out and log in again.
Now, install the cuda toolkit and ati stream.
Add a .conf file into /etc/ld.so.conf.d/ containing the locations of the cuda libraries (typically /usr/local/cuda/lib), run ldconfig.
Logout as root.
Now login as a regular user (at runlevel 5), set up the path, manpath and environment variables appropriately, and try it!
(Hope I haven't forgotten anything...!).
Best,
Steven.
Hello Steven,
Thank you for your hardwork and time. This information would be VERY useful!!!!
Great!
THANKS & Best Regards,
Sarnath
Hi Steven,
It's wonderful info.
Greatly thanks.
Regards,
Seiya
Thanks for the detailed howto, Steven.
I was planning a 4 x 4870x2 system, but it appears the Linux 9.5 fglrx drivers are limited to a max of 4 GPUs. In light of what you are telling me, I'm compelled to put in 2 gtx295 to fill my 4 double wide slots, and have a tag-team race.
I guess it should not be a problem to mix cuda and cal threads.