lebeg opened a new issue #11565: Using ThreadedEnginePerDevice crashes on Raspberry Pi URL: https://github.com/apache/incubator-mxnet/issues/11565 ## Description The failure is not deterministic and the library import to python passes in 1 of 4 times. The creation of NDArray's in this case fails though in all cases with similar message that the stack is corrupted. ## Environment info (Required) ``` pi@raspberrypi:~ $ cat /etc/os-release PRETTY_NAME="Raspbian GNU/Linux 8 (jessie)" NAME="Raspbian GNU/Linux" VERSION_ID="8" VERSION="8 (jessie)" ID=raspbian ID_LIKE=debian HOME_URL="http://www.raspbian.org/" SUPPORT_URL="http://www.raspbian.org/RaspbianForums" BUG_REPORT_URL="http://www.raspbian.org/RaspbianBugs" pi@raspberrypi:~ $ /lib/arm-linux-gnueabihf/libc.so.6 --version GNU C Library (Debian GLIBC 2.19-18+deb8u10) stable release version 2.19, by Roland McGrath et al. Copyright (C) 2014 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Compiled by GNU CC version 4.8.4. Compiled on a Linux 3.16.7 system on 2017-06-19. Available extensions: crypt add-on version 2.1 by Michael Glad and others GNU Libidn by Simon Josefsson Native POSIX Threads Library by Ulrich Drepper et al BIND-8.2.3-T5B libc ABIs: UNIQUE For bug reporting instructions, please see: <http://www.debian.org/Bugs/>. ``` ## Error Message: ``` pi@raspberrypi:~ $ python Python 2.7.9 (default, Sep 17 2016, 20:26:04) [GCC 4.9.2] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import mxnet as mx Segmentation fault: 11 Stack trace returned 3 entries: [bt] (0) /usr/local/lib/python2.7/dist-packages/mxnet/libmxnet.so(dmlc::StackTrace()+0x21) [0x7048475a] [bt] (1) /usr/local/lib/python2.7/dist-packages/mxnet/libmxnet.so(+0x413602) [0x7052b602] [bt] (2) /lib/arm-linux-gnueabihf/libc.so.6(__default_sa_restorer_v2+0) [0x76d75180] terminate called after throwing an instance of 'std::system_error' what(): Resource deadlock avoided Aborted ``` ## Steps to reproduce 1. `ci/build.py -p armv7` 2. `pip install mxnet-1.3.0-py2.py3-none-any.whl` on RPi ## What have you tried to solve it? 1. make / cmake based builds 2. using a different RaspberryPi - on some devices it works 3. reflashed the device 4. compiled and linked with openblas statically 5. connected device to more powerful usb-power source 6. python 2/3 MXNet is working properly if an environment variable is set: ``` export MXNET_ENGINE_TYPE="NaiveEngine" ``` By default the engine type is ThreadedEnginePerDevice and this fails immediately if set explicitly.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
