libOpenCL.so is opened in BOINC using dlopen(). I wonder if dlopen() bypasses ld-linux.so? Its man page claims it checks ld.so.cache, which would indicate that it is using ld-linux.so methods. (libOpenCL.so is present in ld.so.cache.)
On Wed, Sep 3, 2014 at 3:30 PM, Rom Walton <[email protected]> wrote: > is BOINC referencing a different dynamic linker than clinfo? > > > > ----- Rom > > > > *From:* [email protected] [mailto:[email protected]] *On Behalf Of *Eric > J Korpela > *Sent:* Wednesday, September 03, 2014 5:52 PM > > *To:* Rom Walton > *Cc:* [email protected]; David Anderson > *Subject:* Re: BOINC can't detect OpenCL under CUDA 6.5 on linux > > > > I don't even see any indication that boinc is attempting to open an opencl > library before pronouncing that there isn't one. stderrgpudetect.txt and > stdoutgpudetect.txt are empty. > > From an strace of boinc --detect_gpus here are the directories searched... > > open("/lib64/tls/libOpenCL.so", O_RDONLY) = -1 ENOENT (No such file or > directory) > open("/lib64/libOpenCL.so", O_RDONLY) = -1 ENOENT (No such file or > directory) > open("/usr/lib64/tls/libOpenCL.so", O_RDONLY) = -1 ENOENT (No such file or > directory) > open("/usr/lib64/libOpenCL.so", O_RDONLY) = -1 ENOENT (No such file or > directory) > > From cat /etc/ld.so.conf.d/*.conf, here's what should have been > searched... > /usr/lib64/atlas > /usr/lib64/ctapi > /usr/lib/llvm > /usr/lib64/llvm > /usr/lib64/mysql > /usr/lib64/nvidia > /usr/lib/nvidia > /usr/lib64/qt-3.3/lib > /usr/lib/wine/ > /usr/lib64/xulrunner > and the defaults (/usr/lib64,/lib64) > > Adding /usr/lib64 to LD_LIBRARY_PATH solved the problem, but shouldn't > have been necessary. > > > > > > On Wed, Sep 3, 2014 at 1:59 PM, Eric J Korpela <[email protected]> > wrote: > > /usr/lib64/nvidia is in the search path through > /etc/ld.so.conf.d/nvidia-lib64.conf (which contains /usr/lib64/nvidia ) > > And the /etc/OpenCL/vendors directory is there... > % cat /etc/OpenCL/vendors/nvidia.icd > libnvidia-opencl.so.1 > > libnvidia-opencl.so.1 is also located in /usr/lib64/nvidia. > > > > > > On Wed, Sep 3, 2014 at 1:34 PM, Rom Walton <[email protected]> wrote: > > 'nvidia' subdirectory? > > > > Is the nvidia subdirectory in the search path? > > > > There should also be stuff in the /etc/OpenCL/vendors directory. > > > > http://wiki.tiker.net/OpenCLHowTo > > > > ----- Rom > > > > *From:* [email protected] [mailto:[email protected]] *On Behalf Of *Eric > J Korpela > *Sent:* Wednesday, September 03, 2014 4:30 PM > > > *To:* Rom Walton > *Cc:* [email protected]; David Anderson > *Subject:* Re: BOINC can't detect OpenCL under CUDA 6.5 on linux > > > > That shouldn't be it... > > % ls -l /usr/lib64/nvidia/libOpenCL.so* > lrwxrwxrwx 1 root root 18 Sep 3 09:34 /usr/lib64/nvidia/libOpenCL.so > -> libOpenCL.so.1.0.0 > lrwxrwxrwx 1 root root 18 Sep 3 09:34 /usr/lib64/nvidia/libOpenCL.so.1 > -> libOpenCL.so.1.0.0 > -rwxr-xr-x 1 root root 21712 Jul 31 20:10 > /usr/lib64/nvidia/libOpenCL.so.1.0.0 > > > > On Wed, Sep 3, 2014 at 1:24 PM, Rom Walton <[email protected]> wrote: > > No, maybe a case problem? > > > > We are looking for "libOpenCL.so". > > > > Line 181 of client/gpu_opencl.cpp. > > > > ----- Rom > > > > *From:* [email protected] [mailto:[email protected]] *On Behalf Of *Eric > J Korpela > *Sent:* Wednesday, September 03, 2014 4:21 PM > > > *To:* Rom Walton > *Cc:* [email protected]; David Anderson > *Subject:* Re: BOINC can't detect OpenCL under CUDA 6.5 on linux > > > > I don't think that's it. The driver is definitely installed. I can run > OpenCL code and CUDA code from the command line. BOINC runs CUDA apps > fine. BOINC just thinks it can't find libOpenCL.so. Is BOINC using static > paths for libOpenCL, rather than letting ld.so find it? > > > > On Wed, Sep 3, 2014 at 1:16 PM, Rom Walton <[email protected]> wrote: > > I found this post: > > http://scientificlinuxforum.org/index.php?showtopic=2666 > > > > I suspect the packages have been reshuffled. > > > > ----- Rom > > > > *From:* [email protected] [mailto:[email protected]] *On Behalf Of *Eric > J Korpela > *Sent:* Wednesday, September 03, 2014 4:11 PM > *To:* Rom Walton > *Cc:* [email protected]; David Anderson > *Subject:* Re: BOINC can't detect OpenCL under CUDA 6.5 on linux > > > > No, there is no libopencl in /usr/lib or /usr/lib64. This machine has > never has and devkit other than nvidia's installed on it. > > > > On Wed, Sep 3, 2014 at 1:06 PM, Rom Walton <[email protected]> wrote: > > Is there an libopencl.so in /usr/lib? > > > > If so, was the ICD installed? > > > > > http://www.khronos.org/news/permalink/opencl-installable-client-driver-icd-loader > > > > ----- Rom > > > > *From:* [email protected] [mailto:[email protected]] *On Behalf Of *Eric > J Korpela > *Sent:* Wednesday, September 03, 2014 4:00 PM > *To:* [email protected]; David Anderson; Rom Walton > *Subject:* BOINC can't detect OpenCL under CUDA 6.5 on linux > > > > Following installation of CUDA 6.5 on a RHEL6.5 derivative (Scientific > Linux), BOINC can no longer detect OpenCL Libraries. > > 03-Sep-2014 12:45:09 [---] Starting BOINC client version 7.2.33 for > x86_64-pc-linux-gnu > 03-Sep-2014 12:45:09 [---] log flags: file_xfer, sched_ops, task, > coproc_debug > 03-Sep-2014 12:45:09 [---] Libraries: libcurl/7.19.7 NSS/3.15.3 zlib/1.2.3 > libidn/1.18 libssh2/1.4.2 > 03-Sep-2014 12:45:09 [---] Data directory: /var/lib/boinc > 03-Sep-2014 12:45:09 [---] [coproc] launching child process at > /usr/bin/boinc_client > 03-Sep-2014 12:45:09 [---] [coproc] relative to directory /var/lib/boinc > 03-Sep-2014 12:45:09 [---] [coproc] with data directory /var/lib/boinc > 03-Sep-2014 12:45:09 [---] CUDA: NVIDIA GPU 0: GeForce GT 620 (driver > version unknown, CUDA version 6.5, compute capability 2.1, 2048MB, 2032MB > available, 269 GFLOPS peak) > 03-Sep-2014 12:45:09 [---] CUDA: NVIDIA GPU 1: GeForce GT 520 (driver > version unknown, CUDA version 6.5, compute capability 2.1, 1023MB, 1012MB > available, 156 GFLOPS peak) > 03-Sep-2014 12:45:09 [---] NVIDIA library reports 2 GPUs > 03-Sep-2014 12:45:09 [---] No ATI library found > 03-Sep-2014 12:45:09 [---] No OpenCL library found > 03-Sep-2014 12:45:09 [---] App version needs OpenCL but GPU doesn't > support it > 03-Sep-2014 12:45:09 [SETI@home] Application uses missing NVIDIA GPU > 03-Sep-2014 12:45:09 [---] App version needs OpenCL but GPU doesn't > support it > 03-Sep-2014 12:45:09 [Milkyway@Home] Application uses missing NVIDIA GPU > 03-Sep-2014 12:45:09 [---] App version needs OpenCL but GPU doesn't > support it > 03-Sep-2014 12:45:09 [Milkyway@Home] Application uses missing NVIDIA GPU > 03-Sep-2014 12:45:09 [Milkyway@Home] Missing coprocessor for task > ps_modfit_15_3s_130_wrap_const_1_1405680903_8114046_0 > > [many deleted] > > 03-Sep-2014 12:45:09 [SETI@home Beta Test] Missing coprocessor for task > ap_30my13ac_B3_P0_00026_20140825_22705.wu_1 > 03-Sep-2014 12:45:09 [---] Host name: backer > 03-Sep-2014 12:45:09 [---] Processor: 8 GenuineIntel Intel(R) Xeon(R) CPU > E5-2603 v2 @ 1.80GHz [Family 6 Model 62 Stepping 4] > 03-Sep-2014 12:45:09 [---] Processor features: fpu vme de pse tsc msr pae > mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse > sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon > pebs bts rep_good xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 > monitor ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm pcid dca sse4_1 sse4_2 > x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm arat epb > xsaveopt pln pts dts tpr_shadow vnmi flexpriority ept vpid fsgsbase smep > erms > 03-Sep-2014 12:45:09 [---] OS: Linux: 2.6.32-431.23.3.el6.x86_64 > 03-Sep-2014 12:45:09 [---] Memory: 126.00 GB physical, 128.00 GB virtual > 03-Sep-2014 12:45:09 [---] Disk: 78.74 GB total, 56.99 GB free > 03-Sep-2014 12:45:09 [---] Local time is UTC -7 hours > 03-Sep-2014 12:45:09 [---] VirtualBox version: 4.1.34r95024 > 03-Sep-2014 12:45:09 [---] Config: use all coprocessors > > [more deleted] > > 03-Sep-2014 12:45:09 [---] General prefs: from http://bam.boincstats.com/ > (last modified 02-Jul-2014 09:28:55) > 03-Sep-2014 12:45:09 [---] Host location: none > 03-Sep-2014 12:45:09 [---] General prefs: using your defaults > 03-Sep-2014 12:45:09 [---] Reading preferences override file > 03-Sep-2014 12:45:09 [---] Preferences: > 03-Sep-2014 12:45:09 [---] max memory usage when active: 42579.43MB > 03-Sep-2014 12:45:09 [---] max memory usage when idle: 129028.59MB > 03-Sep-2014 12:45:09 [---] max disk usage: 64.78GB > 03-Sep-2014 12:45:09 [---] max CPUs used: 4 > 03-Sep-2014 12:45:09 [---] don't compute while active > 03-Sep-2014 12:45:09 [---] don't use GPU while active > 03-Sep-2014 12:45:09 [---] (to change preferences, visit a project web > site or select Preferences in the Manager) > 03-Sep-2014 12:45:09 [---] Not using a proxy > 03-Sep-2014 12:45:11 Initialization completed > 03-Sep-2014 12:45:11 [Einstein@Home] [coproc] Assigning NVIDIA instance 0 > to PB0037_01111_32_1 > 03-Sep-2014 12:45:11 [Einstein@Home] [coproc] Assigning NVIDIA instance 1 > to PB0037_009D1_126_1 > > clinfo doesn't have any problem finding them. > > Number of platforms: 1 > Platform Profile: FULL_PROFILE > Platform Version: OpenCL 1.1 CUDA 6.5.14 > Platform Name: NVIDIA CUDA > Platform Vendor: NVIDIA Corporation > Platform Extensions: > cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing > cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll > > > Platform Name: NVIDIA CUDA > Number of devices: 2 > Device Type: CL_DEVICE_TYPE_GPU > Device ID: 4318 > Max compute units: 2 > Max work items dimensions: 3 > Max work items[0]: 1024 > Max work items[1]: 1024 > Max work items[2]: 64 > Max work group size: 1024 > Preferred vector width char: 1 > Preferred vector width short: 1 > Preferred vector width int: 1 > Preferred vector width long: 1 > Preferred vector width float: 1 > Preferred vector width double: 1 > Native vector width char: 1 > Native vector width short: 1 > Native vector width int: 1 > Native vector width long: 1 > Native vector width float: 1 > Native vector width double: 1 > Max clock frequency: 1400Mhz > Address bits: 32 > Max memory allocation: 536788992 > Image support: Yes > Max number of images read arguments: 128 > Max number of images write arguments: 8 > Max image 2D width: 32768 > Max image 2D height: 32768 > Max image 3D width: 2048 > Max image 3D height: 2048 > Max image 3D depth: 2048 > Max samplers within kernel: 16 > Max size of kernel argument: 4352 > Alignment (bits) of base address: 4096 > Minimum alignment (bytes) for any datatype: 128 > Single precision floating point capability > Denorms: Yes > Quiet NaNs: Yes > Round to nearest even: Yes > Round to zero: Yes > Round to +ve and infinity: Yes > IEEE754-2008 fused multiply-add: Yes > Cache type: Read/Write > Cache line size: 128 > Cache size: 32768 > Global memory size: 2147155968 > Constant buffer size: 65536 > Max number of constant args: 9 > Local memory type: Scratchpad > Local memory size: 49151 > Kernel Preferred work group size multiple: 32 > Error correction support: 0 > Unified memory for Host and Device: 0 > Profiling timer resolution: 1000 > Device endianess: Little > Available: Yes > Compiler available: Yes > Execution capabilities: > Execute OpenCL kernels: Yes > Execute native function: No > Queue properties: > Out-of-Order: Yes > Profiling : Yes > Platform ID: 0xc42ef0 > Name: GeForce GT 620 > Vendor: NVIDIA Corporation > Device OpenCL C version: OpenCL C 1.1 > Driver version: 340.29 > Profile: FULL_PROFILE > Version: OpenCL 1.1 CUDA > Extensions: > cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing > cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll > cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics > cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics > cl_khr_fp64 > > > Device Type: CL_DEVICE_TYPE_GPU > Device ID: 4318 > Max compute units: 1 > Max work items dimensions: 3 > Max work items[0]: 1024 > Max work items[1]: 1024 > Max work items[2]: 64 > Max work group size: 1024 > Preferred vector width char: 1 > Preferred vector width short: 1 > Preferred vector width int: 1 > Preferred vector width long: 1 > Preferred vector width float: 1 > Preferred vector width double: 1 > Native vector width char: 1 > Native vector width short: 1 > Native vector width int: 1 > Native vector width long: 1 > Native vector width float: 1 > Native vector width double: 1 > Max clock frequency: 1620Mhz > Address bits: 32 > Max memory allocation: 268222464 > Image support: Yes > Max number of images read arguments: 128 > Max number of images write arguments: 8 > Max image 2D width: 32768 > Max image 2D height: 32768 > Max image 3D width: 2048 > Max image 3D height: 2048 > Max image 3D depth: 2048 > Max samplers within kernel: 16 > Max size of kernel argument: 4352 > Alignment (bits) of base address: 4096 > Minimum alignment (bytes) for any datatype: 128 > Single precision floating point capability > Denorms: Yes > Quiet NaNs: Yes > Round to nearest even: Yes > Round to zero: Yes > Round to +ve and infinity: Yes > IEEE754-2008 fused multiply-add: Yes > Cache type: Read/Write > Cache line size: 128 > Cache size: 16384 > Global memory size: 1072889856 > Constant buffer size: 65536 > Max number of constant args: 9 > Local memory type: Scratchpad > Local memory size: 49151 > Kernel Preferred work group size multiple: 32 > Error correction support: 0 > Unified memory for Host and Device: 0 > Profiling timer resolution: 1000 > Device endianess: Little > Available: Yes > Compiler available: Yes > Execution capabilities: > Execute OpenCL kernels: Yes > Execute native function: No > Queue properties: > Out-of-Order: Yes > Profiling : Yes > Platform ID: 0xc42ef0 > Name: GeForce GT 520 > Vendor: NVIDIA Corporation > Device OpenCL C version: OpenCL C 1.1 > Driver version: 340.29 > Profile: FULL_PROFILE > Version: OpenCL 1.1 CUDA > Extensions: > cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing > cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll > cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics > cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics > cl_khr_fp64 > > The libraries exist... > % ldd /usr/bin/clinfo > /usr/bin/clinfo: /usr/lib64/nvidia/libOpenCL.so.1: no version information > available (required by /usr/bin/clinfo) > linux-vdso.so.1 => (0x00007fff199ff000) > libOpenCL.so.1 => /usr/lib64/nvidia/libOpenCL.so.1 > (0x00007feb42148000) > libpthread.so.0 => /lib64/libpthread.so.0 (0x0000003aa9400000) > libstdc++.so.6 => /usr/lib64/libstdc++.so.6 (0x00000039be800000) > libm.so.6 => /lib64/libm.so.6 (0x00000039b4800000) > libdl.so.2 => /lib64/libdl.so.2 (0x00000039b5000000) > libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00000039be400000) > libc.so.6 => /lib64/libc.so.6 (0x00000039b4400000) > /lib64/ld-linux-x86-64.so.2 (0x00000039b4000000) > > > > > > > > > > > _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
