Hardware Acceleration WG, meeting #6

[vmayoral]

Hello ROSers,

I’m calling for the sixth Hardware Acceleration Working Group meeting. This time to be held on 2022-02-22T17:00:00Z (your browser’s time). In case you missed past meetings, here’s the ROS 2 Hardware Acceleration Working Group list with past recordings we are keeping. As for the working group resources, find them below:

ROS 2 Hardware Acceleration Working Group (HAWG) resources

• Objective: Drive creation, maintenance and testing of acceleration kernels on top of open standards (C++ and OpenCL) for optimized ROS 2 and Gazebo interactions over different compute substrates (including FPGAs, GPUs and ASICs).
• Reference hardware platform: Kria K26 Adaptive SOM
• Meeting invite group: ROS 2 Hardware Acceleration WG Google Group
• Instant messaging: Matrix community (Matrix is an open network for secure, decentralized communication).
• Backlog management: https://github.com/ros-acceleration/community/projects/1
• Github organization: ros-acceleration
• Meeting minutes: HAWG minutes
• Discourse tag: wg-acceleration
• Previous meetings:
  • ROS 2 Hardware Accelerating WG, meeting #1
  • ROS 2 Hardware Acceleration WG, meeting #2
  • ROS 2 Hardware Acceleration WG, meeting #3
  • ROS 2 Hardware Acceleration WG, meeting #4
  • ROS 2 Hardware Acceleration WG, meeting #5

The meeting will as usual first review the progress of the WG. As a new topic, we’ll be discussing a “Methodology for ROS 2 Hardware Acceleration” that has been tested in various commercial solutions for hardware acceleration (including Xilinx’s and Nvidia’s). The session will also include a guest talk provided by Christian Lienen who is a PhD candidate at Paderborn University. He will tell us about a hardware-accelerated ROS 2 executor. Their novel approach introduces an FPGA-accelerated event-based ROS 2 executor which schedules callbacks of ROS 2 nodes and utilizes a reconfigurable slot model and partial runtime reconfiguration to load hardware-based callbacks on demand. Their executor registers ROS 2 node functions as callbacks and dispatches them to run either in software or in hardware. Join us to learn more!

The meeting initially scheduled to be 60 minutes. Coordinates for the meeting below:

• Time: 2022-02-22T17:00:00Z
• Coordinates:
  Zoom

  • Meeting URL: Zoom URL
  • Phone one-tap: US: +16699006833,95945587731#,0#,7966693861# or +17209289299,95945587731#,0#,7966693861#
  • Meeting ID: 959 4558 7731
  • Passcode: %5Zn?ig8AH

• Preliminary agenda:
  1. Welcome and newcomers intro
  2. Working group progress review:
     • #20 Methodology for ROS 2 Hardware Acceleration
     • #9 Testing and benchmarking methodology for acceleration kernels
     • (other topics)
  3. Guest talk: Christian Lienen, ReconROS Executor: Event-Driven Programming of FPGA-accelerated ROS 2 Applications
  4. Q&A

[vmayoral]

A couple of additional reminders for those following the WG:

Hardware acceleration in ROS 2 and Gazebo survey is still open!

The hardware acceleration survey for the ROS and Gazebo communities is still open. Data is still coming in. Let us know what are your preferences!

Updated list of WG goals for 2022

During last meeting we updated the list of goals for the WG.

Target	Goal
2021	1) Design tools and conventions to seamlessly integrate acceleration kernels and related embedded binaries into the ROS 2 computational graphs leveraging its existing build system (ament_acceleration extensions) [1], meta build tools (colcon-acceleration extension) and a new firmware layer (acceleration_firmware) [2].
2021	2) Provide reference examples and blueprints for acceleration architectures used in ROS 2 and Gazebo.
2022	3) Facilitate testing environments that allow to benchmark accelerators with special focus on power consumption and time spent on computations (see HAWG benchmarking approach, community#9, tracetools_acceleration, ros2_kria)
2022	4) Survey the community interests on acceleration for ROS 2 and Gazebo (see discourse announcement, survey).
2022	5) Produce demonstrators with robot components, real robots and fleets that include acceleration to meet their targets (see acceleration_examples).
2022	7) Acceleration of complete ROS 2 computational graphs
2022	8) Merge first hardware accelerators (kernels) into upstream packages (candidate: image_pipeline, see image_pipeline instrumented at #717 )
2022	9) Documentation and a “methodology to hardware accelerate a ROS 2 package”
2022	10) Organize workshops on robotics and ROS 2 Hardware Acceleration

Complete details available in the community repository of the Hardware Acceleration WG.

---

1. See ament_vitis ↩︎

2. See acceleration_firmware_kv260 for an exemplary vendor extension of the acceleration_firmware package ↩︎

[ggrigor]

@vmayoral

This time to be held on February 22, 2022 9:00 AM (your browser’s time)

@vmayoral

The meeting initially scheduled to be 60 minutes. Coordinates for the meeting below:

• Time: January 25, 2022 10:00 AM

Is the meeting invite correct, as it conflicts with the stated time in the post?

Thanks

[vmayoral]

@ggrigor

Is the meeting invite correct, as it conflicts with the stated time in the post?

You’re totally right . Sorry about that, updated only the first one. Now both updated!

2022-02-22T17:00:00Z it is!

[vmayoral]

Meeting is happening today: 2022-02-22T17:00:00Z!

As a heads up, we’ll be touching on a recent case study accelerating ROS 2 perception and on various acceleration_examples just disclosed:

ROS 2 Node acceleration
publisher_xilinx	simple_adder	vadd_publisher
doublevadd_publisher	offloaded_doublevadd_publisher	accelerated_vadd_publisher
faster_doublevadd_publisher	multiple_doublevadd_publisher	streaming_k2k_mm_xrt
image_proc_adaptive	image_pipeline_examples
ROS 2 Graph acceleration
perception_2nodes

The invited guest talk today’s also going to very intesting. Christian will will discuss his hardware-accelerated ROS 2 executor.

[vmayoral]

Here’s the recording:

Minutes | Slides