TSMC’s roadmap; AI-based video upscaling; mobility startups

The 3-nanometer node is approaching, with foundries and EDA vendors preparing to address the future demanding processes. And while “downscaling” is still a key word in chipmaking, “upscaling” is becoming an increasingly important term in video consumption, designating the conversion of old low-definition video content into HDTV or even 4K. Artificial intelligence can help in this seemingly miraculous task. Completing this week’s roundup, we continue to monitor the Silicon Valley tech environment – beyond chips – with a quick look at some mobility-related startups.

TSMC’s roadmap: from N5 to N3

From August 24th to 26th, TSMC held its 2020 Technology Symposium and Open Innovation Platform (OIP) Ecosystem Forum, this year in a virtual format. On this occasion, the Taiwanese foundry provided an overview of its recent achievements and some insights on its roadmap. TSMC’s 5 nanometer N5 technology entered volume production this year, providing a 15% performance gain or a 30% power reduction, and up to 80% logic density gain over the preceding N7 technology. Building on the original N5, the company plans to ramp an enhanced N5P version in 2021, offering an additional 5% speed gain and 10% power improvement. TSMC also provided a preview of the latest member of the 5nm family – the N4 process. N4 is expected to offer further PPA improvements with reduced mask layers, while leveraging the 5nm design ecosystem. The N4 process is scheduled to start risk production in fourth quarter of 2021, with volume production in 2022. As for the next node, TSMC claimed to be “on track” with the development of its N3 process, expected to offer up to 15% performance gain, up to 30% power reduction, and a logic density gain up to 70% over N5, also thanks to “architectural innovations”. Major EDA vendors are already preparing to support N3: the achievement of TSMC certification for this future process node has recently been announced by Ansys, Cadence and Synopsys.

TSMC also unveiled its N12e process, now in risk production, targeted at edge AI applications; power savings are obtained by using ultra-low leakage (ULL) device and SRAM. As for packaging, the company introduced the “3DFabric” umbrella name covering the whole range of its advanced technologies: System on Integrated Chips (TSMC-SoIC), Chip on Wafer on Substrate (CoWoS), and Integrated Fan-Out (InFO). Stressing its architectural innovations and advanced packaging capabilities, TSMC seems to be implicitly responding to Intel. On occasion of Intel’s Q2’20 earnings webcast last July 23rd, Intel CEO Bob Swan maintained that “while process technology is very important, it is only one of the six technology pillars of innovation that drive differentiation in our products” adding that the other five pillars are “packaging, architecture, memory, interconnect and security/software.”

Arm to keep IoTP and Data units

According to Bloomberg, Arm said it will keep its IoT Platform and Data units as independent businesses, canceling a previous plan to spin off the units to Softbank. Reportedly, Arm has determined that “both IoTP and Data can realize the same benefits as independent operating businesses, each with their own P&L, under the Arm Limited umbrella with less operational disruption.”

AI-based video upscaling: an Arm-powered solution

Upscaling video content from standard definition (e.g. 480p) to high definition (1080p or more) may sound like the miracle of creating information from nothing: for example, replacing a single pixel with four pixels, and guessing how they should differ from one another to recreate the original image. But AI-based video upscaling takes a less miraculous approach: in the words of Nvidia, “Given a low-resolution image, a deep learning model predicts a high-resolution image that would downscale to look like the original, low-resolution image. To predict the upscaled images with high accuracy, a neural network model must be trained on countless images.” Then, real-time video upscaling requires a processor performing the necessary inferences in real-time on the low-definition video input. Nvidia offers this capability in its Shield streaming media player, but other companies are of course pursuing this goal. Imperial Vision Technology (IVT), a company based in Fuzhou, China, has developed its own solution with the help of Arm. According to an Arm blog post, IVT and Arm have partnered to optimize IVT’s super-resolution algorithm to run on Arm’s Ethos N77 and N78 NPU, and they have achieved “the best performance in the market.” The two companies made sure that all the real-time computation happened on NPU, to avoid performance drops if individual operators would instead run on CPU or GPU. The visual performance was improved by training with a big database of different contents and compensating for the quantization loss – since the “super resolution” model runs on NPU in int8 format. According to Arm, the solution achieved the required performance to support super-resolution from 720p or 1080p to 4K in real-time with excellent VMAF (Video Multimethod Assessment Fusion) scores. Next step of the collaboration between Arm and IVT could be the definition of chipset specifications for DTV, STB, and mobile SoCs.

Process of AI-based video restoration and enhancement. Credit: Arm Limited

Silicon Valley mobility-related startups

Not just Tesla: of course, the Fremont-based electric car maker isn’t the only mobility-related company in the San Francisco Bay Area. Let’s take a quick look at some startups in this region, working in both the EV (electric vehicle) and AV (autonomous vehicle) spaces. Amply (Mountain View, CA) provides electric fleet management solutions to address emerging problems such as energy cost. “In a diesel-vehicle fleet world – Amply points out – the cost of the fuel varies by 10-20% per year, in the electric fuel fleet world the cost of electric fuel often varies by 300% (3X) per day.” Amply aims at removing the risk of time-of-use and demand-charge driven pricing variance, as well as the risk for a fleet operator on choosing charging equipment that could become obsolete. Freewire Technologies (San Leandro, CA) builds EV chargers, claiming to increase the power output by 10-20x at a fraction of the time and cost of deploying other fast chargers. Cruise (San Francisco, CA), a GM’s subsidiary, is an autonomous vehicle developer. Gatik (Palo Alto, CA) develops autonomous vehicles for B2B short-haul logistics – the ‘middle mile’, between two fixed locations. Aurora Innovation (Palo Alto and San Francisco, CA) is an autonomous driving developer on which Amazon has recently invested. Amazon is also reportedly in advanced talks to buy self-driving technology developer Zoox (Foster City, CA).

Credit: FreeWire Technologies

This entry was posted on Monday, August 31st, 2020 at 7:01 am. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.