We were using isaac ros nvblox and vslam on our Jetson Orin AGX for 3D reconstruction and mapping when we got this error that made the nodes shut down:
[component_container_mt-3] CUDA error = 2 at ./nvblox_core/nvblox/include/nvblox/core/internal/impl/unified_ptr_impl.h:39 ‘cudaMallocAsync(&cuda_ptr, sizeof(T), cuda_stream)’. Error string: out of memory.
[ERROR] [component_container_mt-3]: process has died [pid 5433, exit code 99, cmd ‘/opt/ros/humble/lib/rclcpp_components/component_container_mt --ros-args --log-level info --ros-args -r __node:=nvblox_container’].
We have set both map_clearing_radius_m and decay_tsdf_rate_hz to 0.0 in order to view the whole 3D map and save the corresponding mesh. Is it possible to perform large scale mapping without running into this error?