white paper

Google develops WebM video decompression hardware IP using High-Level Synthesis

Success Story

Actual use of Catapult HLS by the WebM team in the successful implementation of the G2 VP9 and share results and impressions. VP9 G2 decoder hardware flow shows the hardware, including both the HLS generated and hand-written RTL blocks.

The WebM project defines an open file format designed for the distribution of compressed media content across the web. Google is a major contributor to the WebM project, having recently undertaken the design and development of the first hardware decoder IP for WebM, otherwise known as the VP9 G2 decoder. This royalty-free hardware IP enables companies developing multimedia system-on-chip (SoC) designs to deliver next-generation performance and power efficiency, enabling up to 4K (2160p 60FPS) resolution playback on consumer devices such as smart TVs, tablet computers and mobile telephones, as well as traditional personal computers and laptops. The VP9 G2 IP has been implemented with a completely new hardware architecture, coded and verified primarily in standard C++ and synthesized to register transfer level (RTL) logic for different target technologies and performance points using Siemens EDA, a part of Siemens Digital Industries Software’s Catapult High-Level Synthesis (HLS).

This paper will describe the actual use of Catapult High-Level Synthesis (HLS) by the WebM team in the successful implementation of the G2 VP9 and share results and impressions. It presents the HLS methodology used to develop the VP9 G2 hardware decoder and explains how it supports the goals and strategies of the WebM project. It explains why the HLS approach makes design implementation and verification 50% faster than a traditional RTL design flow, and how it enables design teams with different end products to collaborate and contribute to the same IP.

Share

Related resources

Bosch Visiontec rapidly brings new automotive IP to market using the Catapult HLS Platform
White Paper

Bosch Visiontec rapidly brings new automotive IP to market using the Catapult HLS Platform

The BOSCH® Visiontec team innovates assisted and autonomous driving technology. This team develops

Chips&Media: Design and Verification of Deep Learning Object Detection IP
White Paper

Chips&Media: Design and Verification of Deep Learning Object Detection IP

Chips&Media decided to adopt a new High-Level Synthesis flow to implement their deep learning algorithm. Enabling them now focus on algorithm and architecture design and reduce painful debugging and expensive computing resources.

STMicroelectronics quickly brings automotive image signal processing to market with High-Level Synthesis
White Paper

STMicroelectronics quickly brings automotive image signal processing to market with High-Level Synthesis

STMicroelectronics crafted a unique High-Level Synthesis flow, enabled by templates, to design and verify an image signal processing (ISP) device, fostering getting it to market as fast as possible.