YouTube Designed Its Own Video Transcoding Hardware

This site may earn affiliate commissions from the links on this page. Terms of use.

The rate of video on the internet has been exploding upwards, year afterwards year, as has the number of videos YouTube serves per year. Unfortunately, CPUs and GPUs don't evangelize the kind of yearly performance improvements they once did. Faced with a slowing rate of silicon improvement and rapidly increasing amounts of video, YouTube decided to build its own video transcoding unit, or VCU, codenamed Argos.

The company has disclosed its Argos effort in both a blog post and a newspaper, depending on how deep into the details you feel similar digging. According to YouTube, moving workloads to the VCU has improved efficiency by twenty-33x depending on the exact particulars of the stream. YouTube'southward new bit is designed to exist capable of transcoding to one resolution target at a time, or of targeting multiple resolutions simultaneously.

A key component of YouTube'south power savings is the fact that the software and hardware stacks are explicitly designed to work with each other. The physical compages of the system is shown beneath:

In that location are more than encode than decode cores on each iteration of the ASIC, and more than one ASIC on each VCU card. This solution has been designed for dense scaling. Transcoding a video to multiple output resolutions simultaneously is function of how YouTube achieves its power efficiency improvements, equally information technology "allows efficient sharing of control parameters obtained by analysis of the source (e.g., detection of fades/flashes)," co-ordinate to the company. Handling these transcodes in parallel (MOT) is much preferred to doing them one at a time (SOT), as it avoids redundant decoding. At least some of the claimed power efficiency improvements will come from avoiding redundant piece of work. MOT is generally preferred to SOT, as information technology avoids redundant decodes for the aforementioned group of outputs.

In MOT, the video is decoded in one case, scaled to all target resolutions, and then encoded at all relevant targets. YouTube notes that it also designed the ASIC to be able to process multiple MOTs and SOTs simultaneously to further boost efficiency. The bodily encoder is designed to encode H.264 and VP9 in hardware while searching three reference frames. Information technology has a pipelined architecture, local reference stores for motion estimation, and tin can advance entropy encoding, simply Google notes the fleck is "optimized for ability/performance/area targets." Each encoder core is capable of encoding 4K at 60fps in real time, with 10 cores per ASIC, and multiple ASICs per card.

YouTube is already drawing upwardly plans for a adjacent-generation accelerator that would besides be capable of decoding AV1 in hardware. VP9 is generally considered to be the open-source competitor for HEVC, while AV1 is a more advanced follow-up expected to deliver greater bandwidth savings.

Argos represents the kind of visitor-specific project we've seen more of in contempo years as Intel has struggled to better its CPU performance, simply this isn't strictly a CPU issue. The GPU decode blocks congenital into an Ampere or RDNA2 GPU conspicuously weren't specialized for the task YouTube had in mind. This is the kind of semi-custom piece of work one could theoretically come across AMD taking on, but AMD doesn't announced to accept pursued outside manufacturing deals for its IP all that aggressively. We know the company is working on a deal with Samsung for a mobile graphics solution based on Radeon IP, and it partners with Sony and Microsoft for console gaming, but non much beyond that — at least, not publicly.

Ten years ago, Google, Facebook, and Amazon began to quietly revolutionize the server marketplace by paying ODMs to build servers for them directly rather than buying off-the-shelf standardized hardware from the likes of Dell or HPE. Today, these aforementioned companies are designing their own custom silicon to fill diverse deject industry use-cases. CPUs and GPUs still dominate the consumer space, merely specialized accelerators and purpose-congenital chips are creeping into the enterprise in e'er-increasing numbers. It's also interesting to run into YouTube rather pointedly not backing HEVC or even discussing hereafter back up for VVC / H.266. Whatsoever abstention of these standards would likely exist due to royalty entanglements and licensing fees.

Feature image by YouTube.

At present Read: