[Linux][Optimus] Gamma-ray pulsar binary search #1 on GPUs : Error 209
27 Dec 2016 22:42:32 UTC
Topic 204104
(moderation:
As this is my first message here I'd like to say: Hello!
I have some issues running E@H projects on GPU.
My set-up is a laptop "hyperbook N55" ( Clevo N550RC):
Intel® Core™ i7-6700HQ
NVIDIA® GeForce GTX 950M + Intel® HD Graphics 530
16GB RAM
OpenSuSE Leap 42.2 + Bumblebee
I've tried to run BOINC Client with optirun using several different settings:
optirun <boincclient command>
optirun --no-xorg <boincclient command>
optirun --no-xorg sudo <boincclient command>
optirun -b primusrun <boincclient command>
but each time I run Gamma-ray pulsar binary search #1 GPU WU I'm getting following error:
boinc_get_opencl_ids returned [0x1465fc0 , 0x1465f10]
Using OpenCL platform provided by: NVIDIA Corporation
Using OpenCL device "GeForce GTX 950M" by: NVIDIA Corporation
Max allocation limit: 524910592
Global mem size: 2099642368
Couldn't create OpenCL context (error: 209)!
initialize_ocl returned error [2007] OCL context null OCL queue null
I've tried to search Internet for the meaning of error 209 but I couldn't find anything useful, Does anyone encountered this type of error and maybe has a solution?
and btw. I've got a Gamma-ray pulsar binary search #1GPU WU that says in details tab "Virtual memory size: 28.21GB" - that type of WU is normally generated?
Hello and welcome to the
)
Hello and welcome to the message boards!
That's ~500 MB and the "Gamma-ray pulsar binary search #1 on GPUs" requires 1 GB of GPU RAM with the present version (1.17), sometime in the (hopefully) near future a version with lowered requirements should be released and it will probably require ~768 MB GPU RAM. See if you could allocate some more memory to your GPU in your BIOS/UEFI.
Thanks for feedback! It
)
Thanks for feedback!
It seems that my system wasn't fully configured for running OpenCL: after reading output of strace I found that apps can't find some shared libraries. In my case they were located in /usr/src/NVIDIA-Linux-x86_64-375.26/ . After I added it to LD_LIBRARY_PATH E@H GPU WU seems to run ok. Fortunately there is no need to change max allocation limit for now.