============================================
  Gazebo Hotspot Analysis: moving_robots_and_sensors
============================================
  Total samples: 44558674825

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  1. Optimization targets (inclusive time)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  Gazebo functions ranked by total CPU spent under them
  (including all children). If you optimize or remove one
  of these functions, you'd save up to this % of CPU.
  These are the primary optimization candidates.

  63.8%  gz::sim::v11::SimulationRunner::Run
  63.8%  gz::sim::v11::ServerPrivate::Run
  63.7%  gz::sim::v11::SimulationRunner::Step
  63.5%  gz::sim::v11::SimulationRunner::UpdateSystems
  59.7%  gz::sim::v11::systems::Physics::Update
  49.3%  gz::sim::v11::systems::PhysicsPrivate::Step
  34.1%  gz::sim::v11::systems::SensorsPrivate::RunOnce
  34.1%  gz::sim::v11::systems::SensorsPrivate::RenderThread
  33.6%  gz::sensors::v11::Sensor::Update
  33.1%  gz::sensors::v11::Manager::RunOnce
  27.5%  gz::sensors::v11::RenderingSensor::Render
  14.3%  gz::sensors::v11::RgbdCameraSensor::Update
  10.3%  gz::sensors::v11::CameraSensor::Update
   8.1%  gz::sim::v11::systems::PhysicsPrivate::UpdateSim
   8.1%  gz::rendering::v11::Ogre2RenderTarget::Render

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  2. What Gazebo functions call (attribution)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  For each sample, finds the nearest Gazebo function in the
  stack and shows what external function it called. Format:
  GazeboFunc -> ExternalFunc. Shows where optimizing the
  Gazebo caller would reduce external library cost.

  49.1%  gz::sim::v11::systems::PhysicsPrivate::Step -> gz::physics::ForwardStep::World<...>
   4.3%  gz::sim::v11::systems::PhysicsPrivate::UpdateCollisions -> gz::physics::GetContactsFromLastStepFeature::World<...>
   2.8%  gz::rendering::v11::Ogre2GpuReadbackTicket::DownloadAndMap -> [[vdso]]
   2.4%  gz::rendering::v11::Ogre2GpuReadbackTicket::DownloadAndMap -> [libnvidia-eglcore.so.580.159.03]
   2.4%  gz::rendering::v11::Ogre2RenderTarget::Render -> Ogre::Node::updateFromParentImpl
   1.4%  gz::rendering::v11::Ogre2ThermalCamera::PostRender -> [[vdso]]
   1.3%  gz::rendering::v11::Ogre2DepthCamera::Render -> Ogre::Node::updateFromParentImpl
   1.2%  gz::rendering::v11::Ogre2ThermalCamera::PostRender -> [libnvidia-eglcore.so.580.159.03]
   1.2%  gz::rendering::v11::Ogre2ThermalCamera::Render -> Ogre::Node::updateFromParentImpl
   1.0%  gz::rendering::v11::Ogre2RenderTarget::Render -> Ogre::ForwardClustered::collectLightForSlice
   1.0%  gz::transport::v16::Node::Publisher::Publish -> __memcpy_avx_unaligned_erms
   0.9%  gz::sim::v11::systems::PhysicsPrivate::UpdateCollisions -> std::unordered_map<...>
   0.9%  gz::sim::v11::systems::PhysicsPrivate::UpdatePhysics -> void
   0.9%  gz::rendering::v11::Ogre2RenderTarget::Render -> [libnvidia-eglcore.so.580.159.03]
   0.8%  gz::rendering::v11::Ogre2RenderTarget::Copy -> [libnvidia-eglcore.so.580.159.03]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  3. External costs (context)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  Top CPU consumers in external libraries (DART, ODE, Ogre,
  glibc, etc.). Not directly actionable, but shows where
  the CPU actually executes. See section 2 for which Gazebo
  function is responsible for each.

  49.1%  gz::physics::ForwardStep::World<gz::physics::FeaturePolicy<double,
   6.8%  [libnvidia-eglcore.so.580.159.03]
   4.9%  Ogre::Node::updateFromParentImpl
   4.9%  [[vdso]]
   2.2%  Ogre::ForwardClustered::collectLightForSlice
   1.8%  __memcpy_avx_unaligned_erms
   1.5%  Ogre::Frustum::updateFrustumPlanesImpl
   1.0%  Ogre::Camera::isViewOutOfDate
   0.8%  Ogre::Frustum::updateWorldSpaceCornersImpl
   0.5%  Ogre::RenderSystem::_makeRsProjectionMatrix
   0.3%  Ogre::Matrix4::inverseAffine
   0.3%  Ogre::Node::convertLocalToWorldPosition
   0.2%  [unknown]
   0.2%  Ogre::Frustum::updateFrustumImpl
   0.2%  Ogre::Frustum::calcProjectionParameters

============================================
  Summary
============================================
  How much CPU runs under a Gazebo function in the call stack?
  This is the portion we can influence by changing Gazebo code.

  Under Gazebo code:  98.1%  (optimization opportunity)
  Outside Gazebo:      1.9%  (external threads, drivers, kernel)