AMD launches custom Alveo media accelerators to handle massive amounts of streaming

Advanced Micro Devices announced the AMD Alveo MA35D media accelerator to help datacenters and companies deal with the modern world of video streaming.

The media accelerator cards feature two 5-nanometer custom chips (application specific integrated circuits, or ASICs) on a card to serve as video processing units (VPUs) supporting the AV1 compression standard. They’re built to power a new era of live interactive streaming services at scale.

With over 70% of the global video market being dominated by live content, a new class of low-latency, high-volume interactive streaming applications are emerging such as watch parties, live shopping, online auctions, and social streaming. Part of the goal is to bring a lot of savings to streaming services that incur the costs of all of that activity.

The prior chip was AMD’s first in this category. The group came through Xilinx, which AMD acquired to get its hands on field-programmable gate arrays. But these are custom chips that are targeted precisely at this market, said Sean Gardner, head of video strategy and market development at AMD, in an interview with VentureBeat.

“This is a new world where you have millions of streams, and you need the absolute best economics,” Gardner said.

The Alveo MA35D media accelerator delivers high channel density, with up to 32 1080p60 streams per card, power efficiency and ultra-low-latency performance critical to reducing the skyrocketing infrastructure costs now required for scaling such compute-intensive content delivery.

“So we are focused on the live portion of the streaming market, which is continuing to grow,” Gardner said. “If you look at all the various streaming or social media, all of them include some portion of live and then you look at live events. Video is becoming a significant portion of all parts of our lives.”

When you go to a concert, you see tons of cameras filming and streaming the experience to others. The official video stream will weave in feeds from lots of cameras and then ship it out for distribution to the broader internet.

“That drives what we call a one-to-millions architecture, and it’s very asymmetrical in nature,” Gardner said. “From a network topology and how the networks are set up. But if you look at the bandwidth, it’s much higher on the download side than the uplink.”

And people aren’t just sitting back and consuming that content. They’re actually adding to it and interacting with it and contributing content.

“This drives different requirements,” Gardner said. “As soon as something is interactive, you have a latency constraint. And the ability and the time in which you need to deliver things becomes much more critical and greatly reduced.”

Compared to the previous generation Alveo U30 media accelerator, the Alveo MA35D delivers up to four times higher channel density, four times max lower latency in 4K and 1.8 times greater compression efficiency to achieve the same VMAF score — a common video quality metric.

“We worked closely with our customers and partners to understand not just their technical requirements, but their infrastructure challenges in deploying high-volume, interactive streaming services profitably,” said Dan Gibbons, general manager of AECG Data Center Group at AMD, in a statement. “We developed the Alveo MA35D with an ASIC architecture tailored to meet the bespoke needs of these providers to reduce both capital and operating expenses for delivering immersive experiences to their users and content creators at scale.”

Purpose-built Video Processing Unit

The Alveo MA35D utilizes a purpose-built VPU to accelerate the entire video pipeline. By performing all video processing functions on the VPU, data movement between the CPU and accelerator is minimized, reducing overall latency and maximizing channel density with up to 32x 1080p60, 8x 4Kp60, or 4x 8Kp30 streams per card.

The platform provides ultra-low latency support for the mainstream H.264 and H.265 codecs and features next-generation AV1 transcoder engines delivering up to a 52% reduction in bitrate for bandwidth savings versus a comparable software implementation.

“AMD’s announcement of the new Alveo MA35D add-in card is an exciting advancement of video acceleration for data centers and is an important step in building out a fully-fledged ecosystem to support royalty-free, high-definition video devices, products, and services,” said Matt Frost, Alliance for Open Media Chair, in a statement. “Live streaming providers are looking for higher density, lower power, lower latency AV1 solutions and by addressing these, alliance members such as AMD are helping facilitate AV1 deployment and overall adoption.”

AI-enabled, intelligent video pipeline

The accelerator features an integrated AI processor and dedicated video quality engines designed to improve the quality of experience at reduced bandwidth.

The AI processor evaluates content, frame-by-frame, and dynamically adjusts encoder settings to improve perceived visual quality while minimizing bitrate. Optimization techniques include region-of-interest (ROI) encoding for text and face resolution, artifact detection to correct scenes with high levels of motion and complexity, and content-aware encoding for predictive insights for bitrate optimization.

Cost effective scaling

Scaling high-volume streaming services require maximizing the number of channels per server while minimizing power and bandwidth-per-stream. By delivering up to 32x 1080p60 streams per card at 1 watt per stream, a 1U rack server equipped with 8 cards delivers up to 256 channels to maximize the number of streams per server, rack or data center.

The competition is predominantly GPUs.

The users are generating their own content, and the volume of streams to process goes up massively. So the video conferencing or video collaboration in many scenarios is no longer a standalone application, Gardner said. Applications like telemedicine, Zoom and others are changing the video patterns as well.

“This is where we see blended experiences, where you’re holding watch party,” he said. “For us, this is really about enabling these applications.”

Software dev kit and product availability

The platform is accessible with the AMD Media Acceleration software development kit (SDK), supporting the widely used FFmpeg and Gstreamer video frameworks for ease of development.

Alveo MA35D media accelerators are sampling now with production shipments expected in Q3. To accelerate development, an Early Access Program is available to qualified customers with comprehensive documentation and software tools for architectural exploration.

Gardner said there are about 22 teraflops of intake per card that the device can encode while consuming just one watt per channel. It processes video at high density, low power and low latency. The price is about $1,600 per card. Each card has multiple video encoders.

“We’ve reduced the bandwidth. We’ve also significantly reduced the latency,” Gardner said.

The result is enabling streaming services to save more money without sacrificing quality.

Originally appeared on: TheSpuzz