LATTE ’23

Workshop on Languages, Tools, and Techniques for Accelerator Design


LATTE is a venue for discussion, debate, and brainstorming about language-oriented approaches to hardware acceleration. The focus is on new languages and tools that aim to let domain specialists, not just hardware experts, produce efficient accelerators.

LATTE '23 is co-located with ASPLOS '23. It will happen on Sunday, March 26, 2023 in Vancouver, BC, Canada.

Registration

To attend LATTE, please register for ASPLOS. Choose the option to attend (at least) one day of workshops/tutorials, and check the box for LATTE on the second page of the registration form.

Call for Participation

Submit your 2-page position paper via HotCRP. Important dates:

Overview

Motivation. Hardware acceleration is a key part of combating the stagnation of hardware performance scaling. Implementing accelerators with state-of-the-art hardware design flows, such as traditional HDLs and current HLS tools, remains a specialized task requiring EE training, proprietary toolchains, and extremely slow compile-edit-run cycles. While traditional approaches might be appropriate for developing general-purpose CPUs that will ship millions of units, they are an impediment to popularizing acceleration for the “long tail” of applications that could benefit from special-purpose hardware. With new language designs and new techniques inspired by traditional compilers research, there is an opportunity to turn accelerator construction from a years-long enterprise into a weekend project.

Scope. LATTE is a venue for discussion, debate, and brainstorming at the intersection of hardware acceleration and programming languages research. The focus is on new languages and tools that aim to let domain specialists, not just hardware experts, produce efficient accelerators. A full range of targets are in scope: ASICs (silicon), FPGAs, CGRAs, or future reconfigurable hardware. A wide variety of research topics are in scope including, but not limited to:

LATTE solicits short position papers that need not fit the mold of a traditional publication:

Position Papers

The primary goal of the workshop is to enable discussion. It will accept 2-page position papers. The workshop will allocate short time slots to the papers, each paired with a discussion following SNAPL's discussion format: “table discussion” where small breakout groups will discuss the paper, followed by plenary Q&A.

Position paper submissions will undergo peer review by a program committee of interdisciplinary experts working on both high-level (languages, compilers, drivers) and low-level (circuit optimization, interconnect design) problems in the area.

Formatting. Papers should use the two-column the formatting guidelines for SIGPLAN conferences (the acmart format with the sigplan two-column option) and not exceed 2 pages, excluding references. Review is single-blind, so please include authors' names on the submitted PDF. We provide a LaTeX example that contains the correct formatting.

Paper submission will is via HotCRP. The accepted papers will not be published in a proceeding—PDFs will instead appear on the workshop's website.

Important guidelines. It's standard for papers to start with a general motivation: Moore's Law is doomed; specialized hardware is ascendant; Verilog is hard to use; etc. Please skip this part in your LATTE position paper (and in eventual talks at the workshop). The LATTE audience will already believe these things, so save the space & time and instead focus on your own unique ideas. As much as possible, dispose with the framing and motivation so you can focus on the technical content.

Remember that the goal at LATTE is to stimulate discussion, not to disseminate fully polished results. So don't be afraid to write up half-baked ideas and in-progress work: it's OK if your submission has zero bar charts, for example.

For examples of past position papers, consider looking at the programs for LATTE ’22 and ’21.

Program Committee

Organizing Committee