From: Karl R. <ru...@iu...> - 2021-11-18 15:53:50
|
Hi Sandra, thanks for the clarification. Can you please add -DVIENNACL_DEBUG_ALL to the compiler invocation? You can set this via CMAKE_CXX_FLAGS either via ccmake, cmake-gui or by calling cmake .. -DCMAKE_CXX_FLAGS="-DVIENNACL_DEBUG_ALL" from your build directory (assuming that this is a subfolder, otherwise adjust the path accordingly) Rerunning the examples should then include a lot of diagnostic output to pin this down. Thanks and best regards, Karli On 11/18/21 16:40, SANDRA CATALAN PALLARES wrote: > Hi, > > I checked a couple of examples (apart from clinfo in the previous mail). > They seem to work fine: > > /jbelloch@odroid:~/EQU_FILTERING_sandra/opencl/samples/hello_world_opencl$ > ./hello_world_opencl > Profiling information: > Queued time: 0.141ms > Wait time: 4.29092ms > Run time: 1.85956ms > jbelloch@odroid:~/EQU_FILTERING_sandra/opencl/samples/hello_world_opencl$ cd > ../template/ > jbelloch@odroid:~/EQU_FILTERING_sandra/opencl/samples/template$ ./template > Profiling information: > Queued time: 0.204ms > Wait time: 0.26867ms > Run time: 0.00108ms/ > > > Here is the output for gdb: > > /jbelloch@odroid:~/EQU_FILTERING_sandra/ViennaCL-1.7.1/examples/tutorial$ gdb > ./blas1 > GNU gdb (Ubuntu 7.7-0ubuntu3.1) 7.7 > Copyright (C) 2014 Free Software Foundation, Inc. > License GPLv3+: GNU GPL version 3 or later > <http://gnu.org/licenses/gpl.html <http://gnu.org/licenses/gpl.html>> > This is free software: you are free to change and redistribute it. > There is NO WARRANTY, to the extent permitted by law. Type "show copying" > and "show warranty" for details. > This GDB was configured as "arm-linux-gnueabihf". > Type "show configuration" for configuration details. > For bug reporting instructions, please see: > <http://www.gnu.org/software/gdb/bugs/ > <http://www.gnu.org/software/gdb/bugs/>>. > Find the GDB manual and other documentation resources online at: > <http://www.gnu.org/software/gdb/documentation/ > <http://www.gnu.org/software/gdb/documentation/>>. > For help, type "help". > Type "apropos word" to search for commands related to "word"... > Reading symbols from ./blas1...(no debugging symbols found)...done. > (gdb) r > Starting program: > /home/jbelloch/EQU_FILTERING_sandra/ViennaCL-1.7.1/examples/tutorial/blas1 > terminate called after throwing an instance of > 'viennacl::ocl::unknown_error' > what(): ViennaCL: FATAL ERROR: ViennaCL encountered an unknown > OpenCL error. Most likely your OpenCL SDK or driver is not installed > properly. In some cases, this error is due to an invalid global work > size or several kernel compilation errors. > If you think that this is a bug in ViennaCL, please report it at > vie...@li... > <mailto:vie...@li...> and supply at least the > following information: > * Operating System > * Which OpenCL implementation (AMD, NVIDIA, etc.) > * ViennaCL version > Many thanks in advance! > > Program received signal SIGABRT, Aborted. > __libc_do_syscall () at > ../ports/sysdeps/unix/sysv/linux/arm/libc-do-syscall.S:44 > 44 ../ports/sysdeps/unix/sysv/linux/arm/libc-do-syscall.S: No such > file or directory. > (gdb) bt > #0 __libc_do_syscall () at > ../ports/sysdeps/unix/sysv/linux/arm/libc-do-syscall.S:44 > #1 0xb6db8f0e in __GI_raise (sig=sig@entry=6) at > ../nptl/sysdeps/unix/sysv/linux/raise.c:56 > #2 0xb6dbb766 in __GI_abort () at abort.c:89 > #3 0xb6f517a8 in __gnu_cxx::__verbose_terminate_handler() () > from /usr/lib/arm-linux-gnueabihf/libstdc++.so.6 > #4 0xb6f501c8 in ?? () from /usr/lib/arm-linux-gnueabihf/libstdc++.so.6 > Backtrace stopped: previous frame identical to this frame (corrupt stack?)/ > / > / > Regards, > Sandra > > El jue, 18 nov 2021 a las 16:27, Karl Rupp (<ru...@iu... > <mailto:ru...@iu...>>) escribió: > > Hi, > > thank you for providing the ldd output. > > The suspicious line is the libOpenCL.so in the LDD output. > > libOpenCL.so => > /home/jbelloch/EQU_FILTERING_sandra/opencl/lib/libOpenCL.so > > looks wrong to me. I'm surprised that clinfo doesn't show an output for > libOpenCL.so, but at least there is this: > /usr/lib/arm-linux-gnueabihf/mali-egl/libmali.so > > Can you please verify that the OpenCL installation in > /home/jbelloch/EQU_FILTERING_sandra/ > is working? I don't know what resides in that folder, but it should > have > some samples to check for correct installation. > > Also, can you please run one of the examples from ViennaCL in gdb and > send a stacktrace? That is, run > gdb ./blas1 > and then at the gdb prompt type > r > and then after the crash type > bt > > Best regards, > Karli > > > > On 11/18/21 14:56, SANDRA CATALAN PALLARES wrote: > > Good afternoon, > > > > I could not make work the examples included in ViennaCL yet. > However, I > > compared the result of ldd on one of the ViennaCL examples and > clinfo. > > The number of linked libraries is significantly different and I > wonder > > if any of them should be also linked for the ViennaCL example and > how > > I can add them to the compilation. > > > > The output of ldd is below. > > > > Thanks in advance. > > > > Best regards, > > Sadra > > > > > jbelloch@odroid:~/EQU_FILTERING_sandra/ViennaCL-1.7.1/examples/tutorial$ > > > ldd amg > > libOpenCL.so => > > /home/jbelloch/EQU_FILTERING_sandra/opencl/lib/libOpenCL.so > (0xb6f19000) > > *libstdc++.so.6 => /usr/lib/arm-linux-gnueabihf/libstdc++.so.6 > (0xb6e4c000)* > > * libm.so.6 => /lib/arm-linux-gnueabihf/libm.so.6 (0xb6de0000) > > libgcc_s.so.1 => /lib/arm-linux-gnueabihf/libgcc_s.so.1 > > (0xb6dbf000) > > libc.so.6 => /lib/arm-linux-gnueabihf/libc.so.6 > (0xb6cd8000)* > > */lib/ld-linux-armhf.so.3 (0xb6f28000)* > > > > > jbelloch@odroid:~/EQU_FILTERING_sandra/ViennaCL-1.7.1/examples/tutorial$ > > > ldd ../../../opencl/samples/clinfo/clinfo > > libmali.so => > /usr/lib/arm-linux-gnueabihf/mali-egl/libmali.so > > (0xb5e17000) > > *libstdc++.so.6 => /usr/lib/arm-linux-gnueabihf/libstdc++.so.6 > (0xb5d6c000)* > > *libm.so.6 => /lib/arm-linux-gnueabihf/libm.so.6 (0xb5d00000) > > libgcc_s.so.1 => /lib/arm-linux-gnueabihf/libgcc_s.so.1 > > (0xb5cdf000) > > libc.so.6 => /lib/arm-linux-gnueabihf/libc.so.6 > (0xb5bf8000)* > > * /lib/ld-linux-armhf.so.3 (0xb6f2c000)* > > libX11.so.6 => /usr/lib/arm-linux-gnueabihf/libX11.so.6 > > (0xb5b12000) > > libX11-xcb.so.1 => > /usr/lib/arm-linux-gnueabihf/libX11-xcb.so.1 > > (0xb5b07000) > > libdrm.so.2 => /usr/lib/arm-linux-gnueabihf/libdrm.so.2 > > (0xb5af7000) > > libxcb.so.1 => /usr/lib/arm-linux-gnueabihf/libxcb.so.1 > > (0xb5add000) > > libxcb-render.so.0 => > > /usr/lib/arm-linux-gnueabihf/libxcb-render.so.0 (0xb5acf000) > > libxcb-dri2.so.0 => > > /usr/lib/arm-linux-gnueabihf/libxcb-dri2.so.0 (0xb5ac4000) > > libXext.so.6 => /usr/lib/arm-linux-gnueabihf/libXext.so.6 > > (0xb5ab0000) > > libXau.so.6 => /usr/lib/arm-linux-gnueabihf/libXau.so.6 > > (0xb5aa5000) > > libXfixes.so.3 => > /usr/lib/arm-linux-gnueabihf/libXfixes.so.3 > > (0xb5a99000) > > libXdmcp.so.6 => /usr/lib/arm-linux-gnueabihf/libXdmcp.so.6 > > (0xb5a8d000) > > libXcomposite.so.1 => > > /usr/lib/arm-linux-gnueabihf/libXcomposite.so.1 (0xb5a82000) > > libXrender.so.1 => > /usr/lib/arm-linux-gnueabihf/libXrender.so.1 > > (0xb5a72000) > > libXrandr.so.2 => > /usr/lib/arm-linux-gnueabihf/libXrandr.so.2 > > (0xb5a64000) > > librt.so.1 => /lib/arm-linux-gnueabihf/librt.so.1 > (0xb5a56000) > > libpthread.so.0 => /lib/arm-linux-gnueabihf/libpthread.so.0 > > (0xb5a3b000) > > libdl.so.2 => /lib/arm-linux-gnueabihf/libdl.so.2 > (0xb5a30000) > > > > El lun, 15 nov 2021 a las 15:25, SANDRA CATALAN PALLARES > > (<sca...@uc... <mailto:sca...@uc...> > <mailto:sca...@uc... <mailto:sca...@uc...>>>) escribió: > > > > Dear Karl, > > > > I installed OpenCL in my home directory and rebuilt Vienna CL, > > however, I still get the same error message. > > > > I checked the installation of OpenCL, the hello_world_opencl > example > > seems to work properly: > > > /jbelloch@odroid:~/EQU_FILTERING_sandra/opencl/samples/hello_world_opencl$ > > ./hello_world_opencl > > Profiling information: > > Queued time: 0.173ms > > Wait time: 4.33126ms > > Run time: 1.92701ms/ > > > > And clinfo seems so: > > > /jbelloch@odroid:~/EQU_FILTERING_sandra/opencl/samples/clinfo$ > ./clinfo > > Number of platforms: 1 > > Platform #0 > > Name: ARM Platform > > Version: OpenCL 1.1 > > > > Number of devices of platform 0: 2 > > Device #0 > > Name: Mali-T628 > > Type: GPU > > Version: OpenCL 1.1 > > Global memory size: 1 GB 970 MB 796 kB > > Local memory size: 32 kB > > Max work group size: 256 > > Max work item sizes: (256, 256, 256) > > > > Device #1 > > Name: Mali-T628 > > Type: GPU > > Version: OpenCL 1.1 > > Global memory size: 1 GB 970 MB 796 kB > > Local memory size: 32 kB > > Max work group size: 256 > > Max work item sizes: (256, 256, 256)/ > > > > In order to compile ViennaCL I just updated the cmake line > and the > > environment variables with the current installation paths getting > > the same error as before. Can you give me any hint about the > > "obstacles in the system setup" that you mention in the > previous email? > > > > Thanks for your help. > > > > Cheers, > > Sandra > > > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>> > > Virus-free. www.avast.com <http://www.avast.com> > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>> > > > > > > <#m_7916694491791944663_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> > > > > El lun, 15 nov 2021 a las 11:53, Karl Rupp > (<ru...@iu... <mailto:ru...@iu...> > > <mailto:ru...@iu... > <mailto:ru...@iu...>>>) escribió: > > > > Dear Sandra, > > > > okay, good, having a working clinfo is a good step. :-) > > > > The cmake line looks fine, also the PATH and LD_LIBRARY_PATH > > export are > > fine. Having OpenCL installed in a different home directory > > sounds a bit > > suspicious, though. Is the user account running ViennaCL the > > same as the > > user account running clinfo? > > > > (Note: We ran quite a number of benchmarks on an ODROID-XU3; > > however, I > > recall that there were a few obstacles in the system > setup that > > we had > > to sort out.) > > > > Best regards, > > Karli > > > > > > > > On 11/15/21 09:44, SANDRA CATALAN PALLARES wrote: > > > Dear Karl, > > > > > > The clinfo seems to work well: > > > > //home/badia/opencl/Mali_OpenCL_SDK_v1.1.0/samples/clinfo/clinfo > > > Number of platforms: 1 > > > Platform #0 > > > Name: ARM Platform > > > Version: OpenCL 1.1 > > > > > > Number of devices of platform 0: 2 > > > Device #0 > > > Name: Mali-T628 > > > Type: GPU > > > Version: OpenCL 1.1 > > > Global memory size: 1 GB 970 MB > 796 kB > > > Local memory size: 32 kB > > > Max work group size: 256 > > > Max work item sizes: (256, 256, 256) > > > > > > Device #1 > > > Name: Mali-T628 > > > Type: GPU > > > Version: OpenCL 1.1 > > > Global memory size: 1 GB 970 MB > 796 kB > > > Local memory size: 32 kB > > > Max work group size: 256 > > > Max work item sizes: (256, 256, 256)/ > > > / > > > / > > > Could it be a problem that OpenCL is installed in another > > user's home? I > > > exported the PATH and LD_LIBRARY_PATH variables > accordingly, > > but still > > > got the same error. > > > > > > These are the steps I followed to compile ViennaCL: > > > > > > /2032 cmake > > -DOPENCLROOT=/home/badia/opencl/Mali_OpenCL_SDK_v1.1.0/ > > > > > > -DOPENCL_LIBRARY=/home/badia/opencl/Mali_OpenCL_SDK_v1.1.0/lib/libOpenCL.so > > > > > ENABLE_UBLAS . > > > 2033 make > > > 2034 ls > > > 2035 cd examples/tutorial/ > > > 2036 ls > > > 2037 ./qr > > > 2038 export > > > > > > PATH=/home/badia/opencl/Mali_OpenCL_SDK_v1.1.0/samples/clinfo:$PATH > > > 2039 export > > > > > > LD_LIBRARY_PATH=/home/badia/opencl/Mali_OpenCL_SDK_v1.1.0/lib:$LD_LIBRARY_PATH > > > 2040 ./qr/ > > > / > > > / > > > Thanks for your help. > > > > > > Best regards, > > > Sandra > > > > > > > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon> > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>>> > > > > > Virus-free. www.avast.com <http://www.avast.com> > <http://www.avast.com <http://www.avast.com>> > > > > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link> > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>>> > > > > > > > > > > > <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> > > > > > > El vie, 12 nov 2021 a las 16:24, Karl Rupp > > (<ru...@iu... <mailto:ru...@iu...> > <mailto:ru...@iu... <mailto:ru...@iu...>> > > > <mailto:ru...@iu... > <mailto:ru...@iu...> > > <mailto:ru...@iu... > <mailto:ru...@iu...>>>>) escribió: > > > > > > Hi, > > > > > > this error usually indicates that your OpenCL > > installation isn't > > > working > > > properly. Does clinfo work from the command line? > > > > > > Best regards, > > > Karli > > > > > > > > > On 11/12/21 14:05, SANDRA CATALAN PALLARES wrote: > > > > Good morning, > > > > > > > > I am trying to run the examples included in Vienna, > > but I get the > > > > following error: > > > > > > > > /./qr > > > > terminate called after throwing an instance of > > > > 'viennacl::ocl::unknown_error' > > > > what(): ViennaCL: FATAL ERROR: ViennaCL > > encountered an unknown > > > > OpenCL error. Most likely your OpenCL SDK or > driver is > > not installed > > > > properly. In some cases, this error is due to an > > invalid global work > > > > size or several kernel compilation errors. > > > > If you think that this is a bug in ViennaCL, please > > report it at > > > > vie...@li... > <mailto:vie...@li...> > > <mailto:vie...@li... > <mailto:vie...@li...>> > > > <mailto:vie...@li... > <mailto:vie...@li...> > > <mailto:vie...@li... > <mailto:vie...@li...>>> > > > > <mailto:vie...@li... > <mailto:vie...@li...> > > <mailto:vie...@li... > <mailto:vie...@li...>> > > > <mailto:vie...@li... > <mailto:vie...@li...> > > <mailto:vie...@li... > <mailto:vie...@li...>>>> and supply at > > least > > > the > > > > following information: > > > > * Operating System > > > > * Which OpenCL implementation (AMD, NVIDIA, etc.) > > > > * ViennaCL version > > > > Many thanks in advance! > > > > Aborted/ > > > > > > > > I am compiling the library on an ODROID-XU3, with > > Ubuntu 14.04.1 > > > LTS. > > > > The OpenCL implementation > is Mali_OpenCL_SDK_v1.1.0, > > which seems > > > to work > > > > properly and the Vienna version is 1.7.1. > > > > > > > > Thanks in advance, > > > > Sandra Catalán > > > > -- > > > > > > > > *Sandra Catalán Pallarés* > > > > SecciónDept. Arquitectura de Computadores y > Automática > > > > Facultad de Ciencias Físicas - Desp. 223 > > > > Universidad Complutense de Madrid > > > > 28040 - Madrid (Spain) > > > > (+34) 91 394 *4394 - *sca...@uc... > <mailto:sca...@uc...> > > <mailto:sca...@uc... <mailto:sca...@uc...>> > <mailto:sca...@uc... <mailto:sca...@uc...> > > <mailto:sca...@uc... <mailto:sca...@uc...>>> > > > <mailto:sca...@uc... <mailto:sca...@uc...> > <mailto:sca...@uc... <mailto:sca...@uc...>> > > <mailto:sca...@uc... <mailto:sca...@uc...> > <mailto:sca...@uc... <mailto:sca...@uc...>>>> > > > > > > > > > > > > _______________________________________________ > > > > ViennaCL-support mailing list > > > > Vie...@li... > <mailto:Vie...@li...> > > <mailto:Vie...@li... > <mailto:Vie...@li...>> > > > <mailto:Vie...@li... > <mailto:Vie...@li...> > > <mailto:Vie...@li... > <mailto:Vie...@li...>>> > > > > > > https://lists.sourceforge.net/lists/listinfo/viennacl-support > <https://lists.sourceforge.net/lists/listinfo/viennacl-support> > > > <https://lists.sourceforge.net/lists/listinfo/viennacl-support > <https://lists.sourceforge.net/lists/listinfo/viennacl-support>> > > > > > > <https://lists.sourceforge.net/lists/listinfo/viennacl-support > <https://lists.sourceforge.net/lists/listinfo/viennacl-support> > > > <https://lists.sourceforge.net/lists/listinfo/viennacl-support > <https://lists.sourceforge.net/lists/listinfo/viennacl-support>>> > > > > > > > > > > > > > > > > -- > > > > > > *Sandra Catalán Pallarés* > > > SecciónDept. Arquitectura de Computadores y Automática > > > Facultad de Ciencias Físicas - Desp. 223 > > > Universidad Complutense de Madrid > > > 28040 - Madrid (Spain) > > > (+34) 91 394 *4394 - *sca...@uc... > <mailto:sca...@uc...> > > <mailto:sca...@uc... <mailto:sca...@uc...>> > <mailto:sca...@uc... <mailto:sca...@uc...> > > <mailto:sca...@uc... <mailto:sca...@uc...>>> > > > > > > > > -- > > > > *Sandra Catalán Pallarés* > > SecciónDept. Arquitectura de Computadores y Automática > > Facultad de Ciencias Físicas - Desp. 223 > > Universidad Complutense de Madrid > > 28040 - Madrid (Spain) > > (+34) 91 394 *4394 - *sca...@uc... > <mailto:sca...@uc...> <mailto:sca...@uc... > <mailto:sca...@uc...>> > > > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>> > > Virus-free. www.avast.com <http://www.avast.com> > > > <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>> > > > > > > <#m_7916694491791944663_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> > > > > > > > > -- > > > > *Sandra Catalán Pallarés* > > SecciónDept. Arquitectura de Computadores y Automática > > Facultad de Ciencias Físicas - Desp. 223 > > Universidad Complutense de Madrid > > 28040 - Madrid (Spain) > > (+34) 91 394 *4394 - *sca...@uc... <mailto:sca...@uc...> > <mailto:sca...@uc... <mailto:sca...@uc...>> > > > > -- > > *Sandra Catalán Pallarés* > SecciónDept. Arquitectura de Computadores y Automática > Facultad de Ciencias Físicas - Desp. 223 > Universidad Complutense de Madrid > 28040 - Madrid (Spain) > (+34) 91 394 *4394 - *sca...@uc... <mailto:sca...@uc...> |