News

PRESS RELEASE: Towards robot accelerators, democratizing hardware acceleration in robotics

Source: Acceleration Robotics, July 25, 2022 [1]
Acceleration Robotics and Harvard University will present in Japan an Open Architecture for hardware accelerating robots that helps prototype “robot chips” and make robot computations faster.

Acceleration Robotics and Harvard present collaborative research

Source: HPCwire, July 25, 2022 [2]
Acceleration Robotics is a robotics semiconductor startup based in the Basque Country, Spain. The core idea behind their research is to facilitate the process of using this technology. Instead of building solely on CPUs, with hardware acceleration (through FPGAs or GPUs) roboticists can empower faster robots.

Robotics hiring levels in the tech industry rose in June 2022

Source: Verdict, July 26, 2022 [3]
The proportion of technology and communications companies hiring for robotics related positions rose in June 2022 compared with the equivalent month last year, with 24.4% of the companies included in our analysis recruiting for at least one such position.

Senate advances bill to boost U.S. microchip production

Source: Forbes, July 26, 2022 [4]
Bill would earmark billions in federal funds to boost U.S. semiconductor production amid a nationwide chip shortage. Key Senate vote could pave the way for legislation to pass through chamber and on to House.

Intel, other chipmakers boost lobbying spend to get CHIPS Act passed

Source: The Register, July 21, 2022 [5]
Intel spent $1.75 million on federal lobbying during Q2 of 2022. Company trying to convince US government to pass CHIPS Act. The act would open $52 billion in funding to boost chip production in the US. Semiconductor industry has spent $19.6 million in federal lobbying so far this year.

Chinese CPU maker Loongson develops custom integrated GPU architecture

Source: Tom's Hardware, July 22, 2022 [6]
The 7A2000 is designed for Loongson's 3-series processors that use proprietary LoongArch architecture and therefore require an operating system and software that support this architecture. The company's customers are various government or local authorities-run organizations that see China-developed IP as a huge benefit.

The Business Case For Location Technologies In Warehousing

Source: Forbes, July 21, 2022 [7]
Logistics managers have seen unprecedented growth in order volume. Recent supply chain disruptions have left many warehouse operators unprepared. Various technological solutions are being deployed in warehouses worldwide. The warehouse automation market is particularly lucrative, with the industry seeing mass mobile robot adoption. In addition, warehouse operators have been ramping up digitization efforts.

Why SoCs need NoCs: network on chip and the future of computing

Source: allaboutcircuits, July 26, 2022 [8]
In a complex heterogeneous computing ecosystem, the development of network-on-chip interconnects is essential to the progress of system-on-chip technology.

Accelerating GPU applications with NVIDIA Math Libraries

Source: NVIDIA Technical Blog, July 25, 2022 [9]
There are three main ways to accelerate GPU applications: using libraries, compilers and programming languages. Preprogrammed libraries offer high-quality implementations of functions encountered in compute-intensive applications. The NVIDIA Math Libraries are optimized to make best use of GPU hardware for the greatest performance gain.

Papers

Evaluation of directive-based heterogeneous redundant design approaches for functional safety systems on FPGAs

Source: J-STAGE Home, May 26, 2022 [10]
Field programmable gate arrays (FPGAs) are used in a wide range of application fields. The resource-level approach focuses on diversity in FPGA technology mapping. Heterogeneous redundant circuit design approaches for FPGAs have a high error detection rate.

Exocompilation for productive programming of hardware accelerators

Source: ACM Conferences, June 9, 2022 [11]
Kernel libraries are critical to exploiting accelerators and specialized instructions in many applications. They are commonly coded and optimized by hand, at great expense, in low-level C and assembly. We propose a new programming language, Exo, based on the principle of exocompilation.

Previous Hardware Acceleration in Robotics Newsletters

Past ROS 2 Hardware Acceleration Working Group meetings

  1. PRESS release: Towards robot accelerators, democratizing hardware acceleration in robotics. (2022, July 25). Hardware Acceleration in Robotics. https://news.accelerationrobotics.com/towards-robot-accelerators-democratizing-hardware-acceleration-in-robotics/ ↩︎

  2. Jolly, A. (2022, July 25). Acceleration robotics and Harvard present collaborative research. HPCwire. https://www.hpcwire.com/off-the-wire/acceleration-robotics-and-harvard-present-collaborative-research/ ↩︎

  3. Data Journalism Team. (2022, July 26). Robotics hiring levels in the tech industry rose in June 2022. Verdict. https://www.verdict.co.uk/robotics-hiring-levels-in-the-tech-industry-rose-in-june-2022/ ↩︎

  4. Portefield, C. (2022, July 26). Senate Advances Bill To Boost U.S. Microchip Production. Forbes. https://www.forbes.com/sites/carlieporterfield/2022/07/26/senate-advances-bill-to-boost-us-microchip-production-heres-whats-in-it/?sh=6368da53dca9 ↩︎

  5. Robinson, D. (2022, July 21). Chipmakers boost lobbying spend to get CHIPS act passed. The Register: Enterprise Technology News and Analysis. https://www.theregister.com/2022/07/21/intel_other_chipmakers_boost_lobbying/?utm_source=daily&utm_medium=newsletter&utm_content=article ↩︎

  6. Shilov, A. (2022, July 22). Chinese CPU maker Loongson develops custom integrated GPU architecture. Tom's Hardware. https://www.tomshardware.com/news/loongson-develops-integrated-gpu-with-2004-functionality ↩︎

  7. Bonte, D. (2022, July 21). The Business Case For Location Technologies In Warehousing. Forbes. https://www.forbes.com/sites/forbestechcouncil/2022/07/21/on-location-the-business-case-for-location-technologies-in-warehousing/?sh=5b8f80b1363e ↩︎

  8. Hertz, J. (2022, July 26). Why SoCs need NoCs: network on chip and the future of computing. allaboutcircuits. https://www.allaboutcircuits.com/news/why-socs-need-nocs-network-on-chip-and-future-computing/ ↩︎

  9. Hunjhunwala, A. J., & Talavera, G. (2022, July 25). Accelerating GPU applications with NVIDIA math libraries. NVIDIA Technical Blog. https://developer.nvidia.com/blog/accelerating-gpu-applications-with-nvidia-math-libraries/ ↩︎

  10. Saikai, T., Miyata, K., Manabe, T., & Shibata, Y. (2022, May 26). Evaluation of directive-based heterogeneous redundant design approaches for functional safety systems on FPGAs. J-STAGE Home. https://www.jstage.jst.go.jp/article/ijnc/12/2/12_387/_article ↩︎

  11. Ikarashi, Y., Bernstein, G. L., Reinking, A., Genc, H., & Ragan-Kelley, J. (2022, June 9). Exocompilation for productive programming of hardware accelerators. ACM Conferences. https://dl.acm.org/doi/10.1145/3519939.3523446 ↩︎