Higher-Order Functions are are a common and convenient way to encapsulate design patterns in software development. However they are not available readily available in hardware design tools. This is because they rely on memory allocators to implement dynamic lists, polymorphism, and looping. These concepts are not very amenable to hardware synthesis.
Dustin presented a solution to this problem in the paper “Higher-Order Functions for C++ Synthesis” at CODES+ISSS as part of Embedded Systems Week. The project develops a library of higher-order functions for C++ High-Level Synthesis tools. We open-sourced these libraries and we hope that you use them. Check that out on github: github.com/drichmond/hops (hops stands for Higher-Order PatternS and also an ode to one of our favorite beverages.). ESWeek was in Torino, Italy so Dustin also found some free time to visit Mont Blanc (pictured).
Once again we had a good showing at 28th International Conference on Field Programmable Logic and Applications (FPL 2018). FPL is the premier European venue for publishing research results in the field of FPGAs and reconfigurable systems. This year, the conference was held in Dublin Ireland. Michael Barrow made the trip to present our two papers.
Hardware design is not easy. It typically involves writing code in low-level languages like Verilog where you must specify how every operation works at every cycle. Modern processors perform billions of operations per second making this is a very difficult task! Yet, hardware design has become increasingly important and more pervasive with the advent of custom accelerators which are used in phones, cars, and in the cloud. We need more hardware designers, but unfortunately, hardware design is hard.
Dr. Dustin Richmond recently defended his PhD thesis that tackled this problem — increasing the accessibility of hardware development to non-hardware engineers through the use of common parallel patterns. As part of this, Dustin developed RIFFA (abstracting communication patterns)  and created a framework for synthesizing higher-order functions to hardware (abstracting computational patterns) .
As with most PhD students in our research group, Dustin had many side projects to distract him during his PhD career. Dustin played a key role in developing our 3D imaging system for creating 3D scans of Maya archaeological sites. This involved expeditions to Guatemala to scan ancient Maya structures, a run-in with a large black snake, and a publication in “Advances in Archaeological Practice” . Dustin also built the hardware for a high framerate 3D imager in one of our first projects with Cognex . This ultimately helped inform Cognex on how to build this sensor which is now a product. Dustin spent two separate internships at Altera (now Intel) and Xilinx. I’m not entirely sure how he fit all of that into one PhD, but certainly, it is impressive.
While PhD defense is mostly focused on research, it should be noted that Dustin has an equally impressive record with university service and teaching. His efforts to our community have been documented in other posts (CSE award and UCSD Graduate Student Association Awards). As a TA, he took on a major revision of our hardware curriculum in the Wireless Embedded Systems Masters Program. He introduced the Xilinx Pynq platform with a series of labs, lectures, and assignments. For the final project, he organized a hackathon where each group was able to make an impressive project in less than two weeks. We will continue to use this curriculum moving forward in that and other classes.
Dustin will continue on the academic route moving back to the Pacific Northwest to be a postdoctoral scholar with Michael Taylor and Luis Ceze. Look for him on the academic job market in 1-2 years.
Dustin Richmond (center with redish shirt) along with the other PhD award winners.
The Department of Computer Science and Engineering gave Dustin Richmond one of its annual awards for “Excellence in Service/Leadership”. Dustin has been a key leader in departmental activities since the day that he arrived N years ago. He has been an active community leader in our department, interacting with staff, students, and faculty to improve our community by organizing and motivating others to do the same. For example, he was the student chair of the department Graduate Community Council. As part of this, he proposed, designed, and oversaw the remodel of Chez Bob. These changes have transformed the lounge into a common meeting area. For several years he also managed the Graduate Student Association budget, allocating money to ideas and projects that improve the quality of life, evaluating and funding ideas that clearly benefit a broad swath of the department, and helping students make their ideas a reality. Dustin also organized an NSF Graduate Research Fellowship Panel, where current applicants can ask questions, and receive feedback about their essays. He has mentored and encouraged other students in the community and recruited faculty with past NSF application experience to participate. This provides a valuable benefit to incoming graduate students and outgoing undergraduate students in our department. These are just some of the highlights; he has done so much more!
Thanks for all your efforts Dustin. Very often these sorts of things go unnoticed. I’m glad that was not the case here.
Power side-channel attacks are a means to extract privileged information, such as secret cryptographic keys, from computational hardware by measuring the subtle variations in voltage drop during the times when the secret data is being computed upon. This is a remarkably simple and effective way to recover secret information using low-cost test equipment.
Our recent research with computer engineers at UW and architects at UCSB brought together hardware design, computer architecture, and statistics to identify and programmatically “blink” the processor when the most information leakage occurs. While blinking, the processor is disconnected from the main power supply and running from an internal capacitor, so that attackers cannot obtain information from measurements of voltage drop during those times. We also explore the trade-offs between area overhead and security, introduce a technique for determining if obvious information leakage exists at processor design time, and a statistical approach to localize this leakage.
“The best book is a finished book” – Anyone that has written a book
Stephen Neuendorffer (Xilinx), Janarbek Matai (Cognex), and I recently “published” the open-source book “Parallel Programming for FPGAs“, which describes how to effectively use high-level synthesis (HLS) tools to program field programmable gate arrays (FPGAs). High-level synthesis is the process of taking an application written in procedural code (e.g., written as C code) and translating it into a hardware design (e.g., like one written in Verilog). HLS tools are seeing a growing usage in the industry as commercial offerings like Xilinx Vivado HLS become more sophisticated. Yet, these tools require the user to understand parallel programming concepts like data partitioning, pipelining, and task level parallelism. These are non-trivial ideas that often are not covered in a programming course. The book aims to clearly explain these concepts while walking through the design and optimization of different applications.
Don’t understand this? Read our book!
The book is open-source with the hope that it will be a living book. We can quickly and easily fix typos, grammar, and poorly written sections. And we can add new materials. You can probably expect at least minor updates during the times when I’m using the book to teach 237C (typically Fall Quarter). We are very open to receiving contributions of new materials, e.g., additions to existing chapters, new chapters, projects, labs, slides, etc.. We would be happy to provide an appropriate level of attribution (e.g., as a chapter author?). Get in touch with me if you are interested in contributing.
Like anything worth doing, this took a lot longer than anticipated (like 5 years longer!), but I am happy to say that we finally got this manuscript into good enough shape to call it version 1.0. Thanks to Steve for making the final push earlier this year. Janarbek was instrumental in providing a lot of the code examples (certainly for all the chapters that I wrote). Much of this came from his PhD thesis and work as a TA for 237C. And a very special thanks to Xilinx for their funding and immense patience on this. They provided no strings attached funding to my lab to allow my PhD students to work on various aspects of this project. I was also able to leverage incredible support through the Wireless Embedded Systems Masters Program (to fund TAs to work specifically on developing the labs and other teaching materials). I know that everyone would have liked this a lot sooner, but I hope that you will agree that this is better late than never?
We have an associated set of projects that we developed alongside this book over the course of many years in my 237C classes. Cleaning up and releasing these projects will likely be the next major revision to the book. I hope to release them broadly this Fall (no promises especially given my track record on getting this book out in a timely manner!). In the meantime, feel free to get in touch with me if you want to have a look at them.
National Geographic announced the Chasing Genius Challenge Finalists which includes undergraduate researcher Nikko Dutra Bouck. Nikko is developing a system that will incentive trash cleanup. It uses low-cost drones to survey an area for trash. Then he plans to develop a Uber/Lyft-like app to pay local people to pick up the trash and deliver it to a landfill. Have a look at his short video for more information and give him a vote while you are there.
The picture shows a large-scale aerial image collected last summer of a mangrove forest in Bahia Magdelena, Mexico. We create this image by stitching together a bunch of pictures that we take from a low-cost drone. It gives us cm-scale resolution which is orders of magnitude better than satellite resolution (like you see on Google Earth). This allows us to zoom in and get fine-grained detail and easily pick out any trash. Even better, we are now working on using automated machine learning algorithms to automatically detect the trash.
Nikko has been working with our research group for more than a year through the Engineers for Exploration program which aims to develop technologies (like this) to aid in conservation, exploration, and cultural heritage. He was awarded an NSF REU scholarship last summer, he is a UCSD FISP awardee, and he leads our collaboration with Octavio Aburto’s group on the mangrove monitoring. He is an amazing person with big goals related to conservation.
Vote now and often and help him take another step towards using the latest technologies to clean up our oceans!
Perry (Lu’s surf instructor :), Ryan, Lu, and the UCSD Hawkinson Bear on Lu’s last day on campus.
Lu Zhang ended his two year stay at UCSD as a visiting graduate student in our group. This was funded by the China Scholarship Council scholarship. During his time at UCSD, he built a measurement system that captured the power consumption of an FPGA while computing different cryptographic operations. His research focused on how different hardware optimizations (e.g., pipelining and memory partitioning) can change the security of the cryptographic hardware. This initial work (“Examining the Consequences of High-Level Synthesis Optimizations on Power Side-Channel”) was accepted for publication in the Design, Automation and Test in Europe (DATE) conference. Lu will travel to Germany to present his work there in mid-March. Lu returns his home university Northwestern Polytechnical University in Xi’an, China to finish his PhD. We wish him the best and look forward to seeing him soon!
University of California Division of War Research illustration of natural underwater sounds, including snapping shrimp, which cause interference with sonar and other underwater acoustic devices. 1944 Scripps Institution of Oceanography Photographs
Snapping shrimp are amazing creatures. They use their claw to create a powerful “sonic boom” that induces a cavitation bubble that stuns their prey. This tiny bubble reaches temperatures of up to 8,000 degrees Fahrenheit and 200 decibels. These little shrimp are so amazing that they were featured on a Radiolab episode.
These shrimp are so loud that they create large disturbances in acoustic field which negatively effects wireless underwater communication and robotic localization; this is illustrated by a 1944 UCSD Scripps Oceanographic cartoon from the UCSD Library Digital Collection). But what if we were able to utilize these snaps in a beneficial way to find the positions of a swarm of underwater vehicles?
Our recent work published in IEEE Access shows how to track a swarm of underwater vehicles using passive signals present in the ocean’s ambient soundscape. We demonstrate our method using noise from these “annoying” snapping shrimp on a swarm of underwater vehicles that was deployed off of San Diego’s coast. Our results show that this works amazingly well. This video shows the results of localization using our snapping shrimp derived localization scheme and the standard, infrastructure-heavy technique of deploying fixed buoys with active acoustic pingers. The latter is one of the most common techniques for localizing underwater systems. Our method is the first step towards an infrastructure free, low power, high endurance localization technique for underwater vehicle swarms.
Neuromorphic vision sensors are uncommon and new(-ish) way to perform analog feature detection in an event-based manner. They employ a fundamentally different technique for sensing compared to “typical” cameras. This results in a more complex pixel detector architecture, and correspondingly means their noise differs from ordinary vision sensors. The picture show KRG PhD student Alireza Khodamoradi recently presented our research in this domain in a paper titled “O(N)-Space Spatiotemporal Filter for Reducing Noise in Neuromorphic Vision Sensors”. Here we analyzed this noise and introduced a novel filter with higher accuracy and minimal memory complexity compared to previous works. This paper was accepted to the IEEE International Conference on Computer Design (ICCD) and was invited to IEEE TETC. This is a special honor only extended to the best papers. We plan to have it published after journal reviews by early 2018.