Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

visualizer segmentation fault #14

Closed
undefinedzero opened this issue Mar 13, 2020 · 5 comments
Closed

visualizer segmentation fault #14

undefinedzero opened this issue Mar 13, 2020 · 5 comments

Comments

@undefinedzero
Copy link

I used the command ./visualizer ~/catkin_suma/src/semantic_suma-master/config/default.xml to run the visualizer, but I encountered a segmentation fault.

OpenGL Context Version 4.6 core profile
GLEW initialized.
OpenGL context version: 4.6
OpenGL vendor string  : NVIDIA Corporation
OpenGL renderer string: GeForce GTX 1080/PCIe/SSE2
Extracting surfel maps partially.
Performing frame-to-model matching.
[1]    21658 segmentation fault (core dumped)  ./visualizer

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7fffc42d1700 (LWP 21782)]
[New Thread 0x7fffb9d9e700 (LWP 21783)]
[New Thread 0x7fffb959d700 (LWP 21784)]
[New Thread 0x7fffb379a700 (LWP 21785)]
OpenGL Context Version 4.6 core profile
GLEW initialized.
OpenGL context version: 4.6
OpenGL vendor string  : NVIDIA Corporation
OpenGL renderer string: GeForce GTX 1080/PCIe/SSE2

Thread 1 "visualizer" received signal SIGSEGV, Segmentation fault.
0x00005555555b2e39 in _mm256_store_ps (__A=..., __P=<optimized out>)
    at /usr/lib/gcc/x86_64-linux-gnu/7/include/avxintrin.h:879
879	  *(__m256 *)__P = __A;
(gdb)

I tried Rangenet, it seems good.

================================================================================
scan: ~/catkin_suma/src/rangenet_lib-master/example/000000.bin
path: ~/catkin_suma/src/semantic_suma-master/darknet53/
verbose: 0
================================================================================
Setting verbosity to: false
Trying to open model
Trying to deserialize previously stored: ~/catkin_suma/src/semantic_suma-master/darknet53//model.trt
Could not deserialize TensorRT engine. 
Generating from sratch... This may take a while...
Trying to generate trt engine from : ~/catkin_suma/src/semantic_suma-master/darknet53//model.onnx
Platform DOESN'T HAVE fp16 support.
No DLA selected.
----------------------------------------------------------------
Input filename:   ~/catkin_suma/src/semantic_suma-master/darknet53//model.onnx
ONNX IR version:  0.0.4
Opset version:    9
Producer name:    pytorch
Producer version: 1.1
Domain:           
Model version:    0
Doc string:       
----------------------------------------------------------------
WARNING: ONNX model has a newer ir_version (0.0.4) than this parser was built against (0.0.3).
 ----- Parsing of ONNX model ~/catkin_suma/src/semantic_suma-master/darknet53//model.onnx is Done ---- 
Success picking up ONNX model
Failure creating engine from ONNX model
Current trial size is 8589934592
Cuda error in file src/implicit_gemm.cu at line 648: out of memory
Success creating engine from ONNX model
Final size is 4294967296
Success creating engine from ONNX model
Trying to serialize engine and save to : ~/catkin_suma/src/semantic_suma-master/darknet53//model.trt for next run
Binding: 0, type: 0
[Dim 5][Dim 64][Dim 2048]
Binding: 1, type: 0
[Dim 20][Dim 64][Dim 2048]
Successfully create binding buffer
================================================================================
Predicting image: ~/catkin_suma/src/rangenet_lib-master/example/000000.bin
================================================================================
Example finished!
@undefinedzero
Copy link
Author

Here are some outputs from valgrind

$ valgrind ./visualizer
==8880== Memcheck, a memory error detector
==8880== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==8880== Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info
==8880== Command: ./visualizer
==8880== 
==8880== Warning: set address range perms: large range [0x16ee1000, 0x2c4fa000) (defined)
OpenGL Context Version 4.6 core profile
GLEW initialized.
OpenGL context version: 4.6
OpenGL vendor string  : NVIDIA Corporation
OpenGL renderer string: GeForce GTX 1080/PCIe/SSE2
here
==8880== Warning: set address range perms: large range [0x59e43040, 0x69e43087) (undefined)
==8880== Warning: set address range perms: large range [0x69e44040, 0x79e44087) (undefined)
Extracting surfel maps partially.
==8880== Warning: set address range perms: large range [0x79e45040, 0x8b0ed840) (undefined)
Performing frame-to-model matching.
==8880== Warning: set address range perms: large range [0x8b0ee040, 0xd2956c87) (undefined)
==8880== Conditional jump or move depends on uninitialised value(s)
==8880==    at 0x530F335: glow::RoSeCamera::matrix() (in /home/bill/catkin_suma/devel/.private/glow/lib/libglow_util.so)
==8880==    by 0x167D39: ViewportWidget::paintGL() (ViewportWidget.cpp:561)
==8880==    by 0x637DCC3: QGLWidget::glDraw() (in /usr/lib/x86_64-linux-gnu/libQt5OpenGL.so.5.9.5)
==8880==    by 0x637D66C: QGLWidget::paintEvent(QPaintEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5OpenGL.so.5.9.5)
==8880==    by 0x6747047: QWidget::event(QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x670883B: QApplicationPrivate::notify_helper(QObject*, QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6710103: QApplication::notify(QObject*, QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x7E0A8D7: QCoreApplication::notifyInternal2(QObject*, QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.9.5)
==8880==    by 0x6740199: QWidgetPrivate::sendPaintEvent(QRegion const&) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6740759: QWidgetPrivate::drawWidget(QPaintDevice*, QRegion const&, QPoint const&, int, QPainter*, QWidgetBackingStore*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6715F5A: QWidgetPrivate::repaint_sys(QRegion const&) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6763D4D: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880== 
==8880== 
==8880== Process terminating with default action of signal 11 (SIGSEGV)
==8880==  General Protection Fault
==8880==    at 0x167D3A: _mm256_load_ps (avxintrin.h:873)
==8880==    by 0x167D3A: pload<__vector(8) float> (PacketMath.h:213)
==8880==    by 0x167D3A: ploadt<__vector(8) float, 32> (GenericPacketMath.h:463)
==8880==    by 0x167D3A: packet<32, __vector(8) float> (CoreEvaluators.h:197)
==8880==    by 0x167D3A: assignPacket<32, 32, __vector(8) float> (AssignEvaluator.h:652)
==8880==    by 0x167D3A: assignPacketByOuterInner<32, 32, __vector(8) float> (AssignEvaluator.h:666)
==8880==    by 0x167D3A: run (AssignEvaluator.h:274)
==8880==    by 0x167D3A: run (AssignEvaluator.h:434)
==8880==    by 0x167D3A: call_dense_assignment_loop<Eigen::Matrix<float, 4, 4>, Eigen::Matrix<float, 4, 4>, Eigen::internal::assign_op<float, float> > (AssignEvaluator.h:741)
==8880==    by 0x167D3A: run (AssignEvaluator.h:879)
==8880==    by 0x167D3A: call_assignment_no_alias<Eigen::Matrix<float, 4, 4>, Eigen::Matrix<float, 4, 4>, Eigen::internal::assign_op<float, float> > (AssignEvaluator.h:836)
==8880==    by 0x167D3A: call_assignment<Eigen::Matrix<float, 4, 4>, Eigen::Matrix<float, 4, 4>, Eigen::internal::assign_op<float, float> > (AssignEvaluator.h:804)
==8880==    by 0x167D3A: call_assignment<Eigen::Matrix<float, 4, 4>, Eigen::Matrix<float, 4, 4> > (AssignEvaluator.h:782)
==8880==    by 0x167D3A: _set<Eigen::Matrix<float, 4, 4> > (PlainObjectBase.h:710)
==8880==    by 0x167D3A: operator= (Matrix.h:208)
==8880==    by 0x167D3A: ViewportWidget::paintGL() (ViewportWidget.cpp:561)
==8880==    by 0x637DCC3: QGLWidget::glDraw() (in /usr/lib/x86_64-linux-gnu/libQt5OpenGL.so.5.9.5)
==8880==    by 0x637D66C: QGLWidget::paintEvent(QPaintEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5OpenGL.so.5.9.5)
==8880==    by 0x6747047: QWidget::event(QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x670883B: QApplicationPrivate::notify_helper(QObject*, QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6710103: QApplication::notify(QObject*, QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x7E0A8D7: QCoreApplication::notifyInternal2(QObject*, QEvent*) (in /usr/lib/x86_64-linux-gnu/libQt5Core.so.5.9.5)
==8880==    by 0x6740199: QWidgetPrivate::sendPaintEvent(QRegion const&) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6740759: QWidgetPrivate::drawWidget(QPaintDevice*, QRegion const&, QPoint const&, int, QPainter*, QWidgetBackingStore*) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6715F5A: QWidgetPrivate::repaint_sys(QRegion const&) (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x6763D4D: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880==    by 0x676466C: ??? (in /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5.9.5)
==8880== 
==8880== HEAP SUMMARY:
==8880==     in use at exit: 2,179,270,026 bytes in 77,052 blocks
==8880==   total heap usage: 326,734 allocs, 249,682 frees, 5,598,045,455 bytes allocated
==8880== 
==8880== LEAK SUMMARY:
==8880==    definitely lost: 512 bytes in 2 blocks
==8880==    indirectly lost: 322 bytes in 13 blocks
==8880==      possibly lost: 1,877,480,202 bytes in 199 blocks
==8880==    still reachable: 301,706,326 bytes in 76,164 blocks
==8880==                       of which reachable via heuristic:
==8880==                         stdstring          : 76,127 bytes in 1,010 blocks
==8880==                         length64           : 4,832 bytes in 80 blocks
==8880==                         newarray           : 2,112 bytes in 52 blocks
==8880==                         multipleinheritance: 808 bytes in 5 blocks
==8880==         suppressed: 0 bytes in 0 blocks
==8880== Rerun with --leak-check=full to see details of leaked memory
==8880== 
==8880== For counts of detected and suppressed errors, rerun with: -v
==8880== Use --track-origins=yes to see where uninitialised values come from
==8880== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
[1]    8880 segmentation fault (core dumped)  valgrind ./visualizer

@jbehley
Copy link
Member

jbehley commented Mar 13, 2020

Problems with Eigen usually originate in clash between gtsam's packaged and the systems Eigen version.

Please try to reinstall gtsam with the compile option USE_SYSTEM_EIGEN set.

@undefinedzero
Copy link
Author

I tried recompiling gtsam with systems Eigen(version 3.3.7), but the problem remains. I fixed this problem by changing the Eigen's version to 3.2.10

@Chen-Xieyuanli
Copy link
Member

@undefinedzero Thank you for the feedback!

@undefinedzero
Copy link
Author

I found that using Eigen 3.2.10 will also encounter a segmentation fault when reading calibration files. I tried 3.2.1, it seems good for now.

@jbehley jbehley mentioned this issue Mar 25, 2020
Closed
@ACFFF ACFFF mentioned this issue Jul 21, 2020
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@jbehley @Chen-Xieyuanli @undefinedzero