Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Anders Damsgaard
    @anders-dc
    test
    This message was deleted
    Anders Damsgaard
    @anders-dc
    test
    test
    this is a test from gitter-irc f(x)=x
    Anders Damsgaard
    @anders-dc
    hlep
    s/help/help/
    Anders Damsgaard
    @anders-dc
    test
    Anders Damsgaard
    @anders-dc
    5/15
    Anshuman Goswami
    @anshumang
    ping
    Anders Damsgaard
    @anders-dc
    @anshumang pong
    @anshumang I'll get back to you by mail later this weekend
    Anshuman Goswami
    @anshumang
    great !
    thanks for looking into this
    Anshuman Goswami
    @anshumang
    @anders-dc just checking back...did you get a chance to look?
    Anders Damsgaard
    @anders-dc
    anshumang: please try the file python shear-test.py. You can increase the number of particles if you want (line 14).
    anshumang: the script first spawns particles randomly in a large space and then lets them fall to the bottom. Afterwards, they are consolidated in a biaxial setup with periodic boundaries under a given normal stress. Finally, they are sheared at the same normal stress at a given shear velocity.
    anshumang: also, apologies for the late reply
    Anshuman Goswami
    @anshumang
    Hi Anders, thanks a lot for following up...I am trying out what you suggested...
    Getting the following error when running shear-test.py -
    NameError: name 'Spherebin' is not defined
    Anders Damsgaard
    @anders-dc
    anshumang: alright, that script was a bit old so it needed a couple of fixes. Please run git pull and try it again
    Anders Damsgaard
    @anders-dc
    anshumang: I can unfortunately not test it on my own system. The GPU's are busy finishing simulations for my thesis
    Anshuman Goswami
    @anshumang
    :+1: works, let you know once I run a CUPTI profile on it to get the kernel running times....
    Anshuman Goswami
    @anshumang
    I see 5 kernels besides the thrust sorting kernels in the CUPTI trace....namely, reorderArrays, topology, interact, integrate, calcParticleCellID...is that what you expect?
    None of these have 3D grids...and all of their running times are less than 1ms...actually, those are two things which I need..
    Any thoughts?....if you are busy with your thesis now, never mind...
    Anders Damsgaard
    @anders-dc
    anshumang: thanks for the feed back! The first part of the model run is going to involve very little grain-interaction, since the particles aren't initially touching. As the granular packing becomes denser and there are more contacts per particle, the topology and interact kernels should become heavier