yade-dev team mailing list archive
-
yade-dev team
-
Mailing list archive
-
Message #14889
[Bug 1852197] Re: Segmentation fault on pfacet gridconnection contact
Ok, found, changed the path to memoizeDb. ASAN says:
=====
==5085==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x7f7d59ac2c6f bp 0x7f7d40cf7ac0 sp 0x7f7d40cf61d0 T16777215)
==5085==The signal is caused by a READ memory access.
==5085==Hint: address points to the zero page.
In [5]: #0 0x7f7d59ac2c6e in yade::Bo1_PFacet_Aabb::go(boost::shared_ptr<yade::Shape> const&, boost::shared_ptr<yade::Bound>&, yade::Se3<double> const&, yade::Body const*) /trunk/pkg/common/PFacet.cpp:817
#1 0x7f7d59141278 in yade::BoundDispatcher::processBody(boost::shared_ptr<yade::Body> const&) /trunk/pkg/common/Dispatching.cpp:54
#2 0x7f7d59142ded in yade::BoundDispatcher::action() [clone ._omp_fn.0] /trunk/pkg/common/Dispatching.cpp:36
#3 0x7f7d50c11e0e in GOMP_parallel (/lib/x86_64-linux-gnu/libgomp.so.1+0xde0e)
#4 0x7f7d591433ee in yade::BoundDispatcher::action() /trunk/pkg/common/Dispatching.cpp:31
#5 0x7f7d59821812 in yade::InsertionSortCollider::action() /trunk/pkg/common/InsertionSortCollider.cpp:309
#6 0x7f7d58c04381 in yade::Scene::moveToNextTimeStep() /trunk/core/Scene.cpp:98
#7 0x7f7d58c12613 in yade::SimulationFlow::singleAction() /trunk/core/SimulationFlow.cpp:26
#8 0x7f7d58ccefc7 in yade::ThreadWorker::callSingleAction() /trunk/core/ThreadWorker.cpp:73
#9 0x7f7d58cc5259 in yade::ThreadRunner::call() /trunk/core/ThreadRunner.cpp:54
#10 0x7f7d58cc71cf in yade::ThreadRunner::run() /trunk/core/ThreadRunner.cpp:28
#11 0x7f7d58cc74c8 in boost::_mfi::mf0<void, yade::ThreadRunner>::operator()(yade::ThreadRunner*) const /usr/include/boost/bind/mem_fn_template.hpp:49
#12 0x7f7d58cc74c8 in void boost::_bi::list1<boost::_bi::value<yade::ThreadRunner*> >::operator()<boost::_mfi::mf0<void, yade::ThreadRunner>, boost::_bi::list0>(boost::_bi::type<void>, boost::_mfi::mf0<void, yade::ThreadRunner>&, boost::_bi::list0&, int) /usr/include/boost/bind/bind.hpp:259
#13 0x7f7d58cc74c8 in boost::_bi::bind_t<void, boost::_mfi::mf0<void, yade::ThreadRunner>, boost::_bi::list1<boost::_bi::value<yade::ThreadRunner*> > >::operator()() /usr/include/boost/bind/bind.hpp:1294
#14 0x7f7d58cc74c8 in boost::detail::function::void_function_obj_invoker0<boost::_bi::bind_t<void, boost::_mfi::mf0<void, yade::ThreadRunner>, boost::_bi::list1<boost::_bi::value<yade::ThreadRunner*> > >, void>::invoke(boost::detail::function::function_buffer&) /usr/include/boost/function/function_template.hpp:159
#15 0x7f7d58cccf6b in boost::function0<void>::operator()() const /usr/include/boost/function/function_template.hpp:768
#16 0x7f7d58cccf6b in boost::detail::thread_data<boost::function0<void> >::run() /usr/include/boost/thread/detail/thread.hpp:117
#17 0x7f7d55a39f64 (/lib/x86_64-linux-gnu/libboost_thread.so.1.67.0+0x14f64)
#18 0x7f7d5f80afa2 in start_thread /build/glibc-vjB4T1/glibc-2.28/nptl/pthread_create.c:486
#19 0x7f7d5f3514ce in clone (/lib/x86_64-linux-gnu/libc.so.6+0xf94ce)
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /trunk/pkg/common/PFacet.cpp:817 in yade::Bo1_PFacet_Aabb::go(boost::shared_ptr<yade::Shape> const&, boost::shared_ptr<yade::Bound>&, yade::Se3<double> const&, yade::Body const*)
==5085==ABORTING
=====
Could you please file a bug on gitlab with this content? Thanks
--
You received this bug notification because you are a member of Yade
developers, which is subscribed to Yade.
https://bugs.launchpad.net/bugs/1852197
Title:
Segmentation fault on pfacet gridconnection contact
Status in Yade:
New
Bug description:
I seem to get a Segmentation fault (core dumped) when using
GridConnection, PFacet, and randomDensePack. And I am not sure why. I
have attached an image and "working" example.
Thanks in advance,
Justin
Edit: I could not seem to upload an picture or example code. Thus code
below:
from yade.gridpfacet import *
import numpy as np
from yade import utils
from yade import ymport
import sys,time,math,random
from yade import qt
qt.View()
#print(os.path.basename(__file__))
#quit()
########### ParallelEngine, -j4
boxCenter = (0,0,.025)
## double check that /2 would create correct length, I.e., import based_70 to see if they match... they should
boxDem = (.2123/2,.212/2,.064/2) ### based_70 is 212.3x212.21x63.65
boxDem = (.07,.07,.025) ###
blen = boxDem[0]*2 ## Length of Ball pit
bhei = .025 ## Height of Ball pit
# Spheres information
sphereRadius = .004 # [m]
nu = .48
#G = 300000 # [Pa]
den_rub = 89724 # [kg/m^3]
yng_rub = 3000000 # [Pa]
fric_rub = radians(38) # [degrees]
O.materials.append(FrictMat(frictionAngle=fric_rub,young=yng_rub,density=den_rub,poisson=nu,label='rubber'))
O.materials.append(FrictMat(young=200e9,density=8050,poisson=.3, label='steel')) # used steel properties to make rigid
## Stud informtion
#studStartZ = (.01+.022+.01905)
studStartZ = (.01+bhei+.01905)
transVel = 0.000222*2
rotVel = 0.00872665*2
### pack auto fills, i.e., I no longer pick the amount of spheres
pred = pack.inAlignedBox((-(boxDem[0]-.001),-(boxDem[0]-.001),bhei+sphereRadius/2+.001),((boxDem[0]-.001),(boxDem[0]-.001),bhei+.045))
sp = pack.randomDensePack(pred, radius=sphereRadius, memoizeDb='tmp/single_grass.sqlite', returnSpherePack=True)
sp.toSimulation(wire=False,material='rubber')
## Time step set to 20% of Rayleigh Wave
O.dt=.2*utils.RayleighWaveTimeStep() # 0.0011302534626965682
print(O.dt)
### Steps needed to run sim
sphereFalls = 1000 ### Steps spheres need to fall and level out and turn on studs
startRot = sphereFalls+math.floor((studStartZ-bhei-.01)/transVel/O.dt) ### Step needed for transVel to get studs level with spheres
## Rotation steps
deg_to_rad_length = (60*pi/180); # [rad]
num_sec_roto = deg_to_rad_length/rotVel; # [s]
num_of_steps_roto = num_sec_roto/O.dt;
endSim = startRot + math.floor(num_of_steps_roto)
### total steps in sim
O.engines=[
###Reset all forces stored in Scene::forces (O.forces in python). Typically, this is the first engine to be run at every step. In addition, reset those energies that should be reset, if energy tracing is enabled.
## Resets forces and moments that act on bodies
ForceResetter(),
## Using bounding boxes find possible body collisions.
InsertionSortCollider([
Bo1_Facet_Aabb(),
Bo1_PFacet_Aabb(),
Bo1_Sphere_Aabb(),
Bo1_GridConnection_Aabb(),
Bo1_PFacet_Aabb(),
]),
InteractionLoop([
Ig2_Facet_Sphere_ScGeom(),
Ig2_Wall_PFacet_ScGeom(),
Ig2_Wall_Sphere_ScGeom(),
Ig2_Sphere_Sphere_ScGeom(),
Ig2_Sphere_GridConnection_ScGridCoGeom(),
Ig2_Sphere_PFacet_ScGridCoGeom(),
Ig2_GridNode_GridNode_GridNodeGeom6D(),
Ig2_GridConnection_GridConnection_GridCoGridCoGeom(),
Ig2_GridConnection_PFacet_ScGeom(),
Ig2_PFacet_PFacet_ScGeom(),
],
[
# Ip2_CohFrictMat_CohFrictMat_CohFrictPhys(setCohesionNow=True,setCohesionOnNewContacts=True), # internal cylinder physics
Ip2_CohFrictMat_CohFrictMat_CohFrictPhys(setCohesionNow=True), # internal cylinder physics
## Currently grass does not interact with facet, probably should not matter, since the spheres will be holding them up.
Ip2_FrictMat_FrictMat_FrictPhys() # physics for external interactions, i.e., cylinder-cylinder, sphere-sphere, cylinder-sphere
],
[
Law2_ScGeom_FrictPhys_CundallStrack(), # contact law for sphere-sphere
Law2_ScGridCoGeom_FrictPhys_CundallStrack(), # contact law for cylinder-sphere
Law2_ScGeom6D_CohFrictPhys_CohesionMoment(), # contact law for "internal" cylinder forces
Law2_GridCoGridCoGeom_FrictPhys_CundallStrack(), # contact law for cylinder-cylinder interaction
]),
NewtonIntegrator(damping=.2,gravity=[0,0,-9.81], label='newtonInt'),
]
## Grass Information
#O.materials.append(CohFrictMat(young=8e5,poisson=0.3,density=4e3,frictionAngle=np.radians(30),normalCohesion=1e5,shearCohesion=1e5,momentRotationLaw=True,label='grass')) ## Properties Need to be corrected
O.materials.append(CohFrictMat(young=100,poisson=0.3,density=40,frictionAngle=np.radians(30),normalCohesion=1e5,shearCohesion=1e5,momentRotationLaw=True,label='grass'))
#rCyl = 0.0006 ## Radius, Grass was about 1.2 [mm] wide
rCyl = 0.0026 ## Grass was about 1.2 [mm] wide
nL = 4 ## No exact Number here, just trial and error
L = bhei ## Height of spheres
### Grass Creation
### Create all nodes first :
nodesIds=[]
idxc = -1
x_gap = 0.009 ## Between lumps is roughly 9 [mm]
y_gap = 0.01905 ## Between lines of backing is .75 inch apart
range_x = int(math.floor(blen/x_gap)) ## finding the range for x
range_y = int(math.floor(blen/y_gap)) ## finding the range for z
cen_y = -(range_y/2)*y_gap ## Allows the "box" of grass to be center in Z
color=[255./255.,102./255.,0./255.]
steel_r = (1/4*0.00635)/2 ### Converting 1/4 steel to m, to a sphere radius
box = []
box.append(O.bodies.append( gridNode([-boxDem[0],-boxDem[1],0],steel_r,wire=False,fixed=True,material='steel',color=color) ))
box.append(O.bodies.append( gridNode([boxDem[0],-boxDem[1],0],steel_r,wire=False,fixed=True,material='steel',color=color) ))
box.append(O.bodies.append( gridNode([boxDem[0],boxDem[1],0],steel_r,wire=False,fixed=True,material='steel',color=color) ))
box.append(O.bodies.append( gridNode([-boxDem[0],boxDem[1],0],steel_r,wire=False,fixed=True,material='steel',color=color) ))
O.bodies.append( gridConnection(box[0],box[1],steel_r,color=color,material='steel') )
O.bodies.append( gridConnection(box[1],box[2],steel_r,color=color,material='steel') )
O.bodies.append( gridConnection(box[2],box[0],steel_r,color=color,material='steel') )
O.bodies.append( gridConnection(box[2],box[3],steel_r,color=color,material='steel') )
O.bodies.append( gridConnection(box[3],box[0],steel_r,color=color,material='steel') )
O.bodies.append( pfacet(box[0],box[1],box[2],wire=False,material='steel',color=color) )
O.bodies.append( pfacet(box[0],box[2],box[3],wire=False,material='steel',color=color) )
O.stopAtIter=endSim ### Stop sim at endSim steps
#O.run()
To manage notifications about this bug go to:
https://bugs.launchpad.net/yade/+bug/1852197/+subscriptions
References
