Highlights from FPGA 2020

John's Blog

Here are a few personal highlights from the FPGA 2020 conference, which took place this week in Seaside, California. (Main photo credit: George Constantinides.)

Jakub Szefer‘s invited talk on “Thermal and Voltage Side Channels and Attacks in Cloud FPGAs” described a rather nifty side-channel through which secrets could be leaked in the context of cloud-based FPGAs. The context is that one user of an FPGA would like to transmit information secretly to the next user of the same FPGA. Jakub suggests transmitting this information via the physical temperature of the FPGA. His proposal is that if the first user would like to transmit a ‘1’, they heat up the FPGA by running a circuit that uses a lot of current, such as a ring oscillator; if they would like to transmit a ‘0’, they don’t. The second user, once they arrive, measures the temperature of…

View original post 1,412 more words

When to Schedule?

On Tuesday, Jianyi Cheng will present our recent work Combining Dynamic and Static Scheduling in High-level Synthesis at the ACM International Symposium on FPGAs in Monterey. This is joint work between Jianyi and his two supervisors, John Wickerson and myself, as well as our collaborators from EPFL, Lana Josipović and her PhD supervisor Paolo Ienne.

As I’ve described in previous blog posts [1,2], Lana has been doing interesting work over the last few years, developing a tool flow for dynamically-scheduled high-level synthesis (HLS). Typically in modern HLS tools like VivadoHLS or LegUp, scheduling decisions are made statically – at compile time. However, Lana’s tool flow makes these decisions dynamically, at run time, using handshaking circuitry, reminiscent of Page and Luk’s work compiling occam into FPGAs.

In our paper, we have teamed up with EPFL to build a flow that can result in the best of both worlds. Static scheduling can be very efficient, sharing resources and leading to low area designs. Dynamic scheduling can be very fast, working around actual rather than potential data dependencies. Jianyi’s paper allows the definition of statically scheduled functions within a dynamically scheduled program. He shows that over a range of benchmarks, the results are about half the area of the fully dynamically-scheduled designs while about 1.7x faster than the fully statically-scheduled designs.

Screenshot 2020-02-21 at 11.07.12