Fab-related updates; Arm-Nvidia deal; HPC and ML rankings

Not surprisingly, several news this week are somewhat related to fabs and foundries – in terms of customer contracts, new fabs, acquisitions, equipment – as companies around the world are repositioning to take advantage of the semiconductor boom. Another interesting update concerns Simon Segars speaking about the Nvidia deal. More news is coming from the International Supercomputing Conference and from the latest MLPerf results.

Apple and Intel reportedly first to adopt TSMC’s 3 nm process

According to Nikkei Asia, Apple and Intel have emerged as the first adopters of TSMC’s 3-nanometer process ahead of its deployment as early as next year. Commercial output of such chips is expected to start in the second half of next year. Apple’s iPad will likely be the first devices powered by processors made using 3-nanometer technology. Intel is reportedly working with TSMC on at least two 3-nanometer projects concerning CPUs for notebooks and data center servers. Mass production of these chips is expected to begin by the end of 2022 at the earliest. Reportedly, the chip volume planned for Intel is more than that for Apple’s iPad.

STMicroelectronics to share a new Italian fab with Tower Semiconductor

STMicroelectronics will share the cleanroom and the facility infrastructure in its Agrate, Italy R3 300mm fab – under construction – with Tower Semiconductor. Tower will install its own equipment in one third of the total space. The goal of the agreement is accelerated fab ramp-up, a key factor to reach a high utilization level and therefore a competitive wafer cost.

ITEC to become an independent semiconductor equipment vendor

Netherlands-based ITEC, a semiconductor equipment manufacturer founded in 1991 by Philips and currently owned by Nexperia, is being launched as a separate independent entity. As stated in a press release from Wingtech Technology – Nexperia’s Chinese parent company – the move “will allow ITEC to address the third-party market in time to serve the current semiconductor boom”. Solutions offered by ITEC include assembly equipment for die bond and chip sort, test platforms, smart vision inspection systems, factory automation software etc.

Credit: Itec

Arm CEO: Nvidia acquisition a better choice than IPO

In a recent blog post, Arm CEO Simon Segars addressed the concerns about the Nvidia-Arm deal and replied to those who would rather prefer Arm to go public. “The combination of Arm and Nvidia is a better outcome than an IPO,” he wrote. “The level of investment that will be needed to lead in AI will be unprecedented. (…) Now is the time for us to take our scale to the next level to address the technology challenges ahead. We contemplated an IPO but determined that the pressure to deliver short-term revenue growth and profitability would suffocate our ability to invest, expand, move fast and innovate. Combining with Nvidia will give us the scale, resources and agility needed to maximize the opportunities ahead. This deal is the best opportunity for Arm and our customers.” Segars continued: “We have built the value of our company through our open-licensing business model which ensures our customers around the world all have access to the same Arm technology. This is an economic and commercial necessity for Arm’s business, and simple common sense. Arm and Nvidia rely heavily on others throughout the industry so continuation of our open licensing model is not just the best option, it is the only one. We’ll invest with Nvidia to create new markets, not displace customers in existing ones.”

Simon Segars. Credit: Arm

Fugaku still first on the Top500 supercomputer list

The Japanese supercomputer Fugaku still holds the top spot on the new edition of the Top500 supercomputer list. Codeveloped by Riken and Fujitsu, Fugaku is based on Fujitsu’s custom Arm A64FX processor and has 7,630,848 cores. It has an HPL benchmark score of 442 Pflop/s, but in single or further reduced precision – which is often used in machine learning and AI – its peak performance is actually above an exaflop. Fugaku in fact reached 2 exaflops on the new HPL-AI benchmark. The rest of the top10 saw little change, too, with only one new entry: the Perlmutter system at NERSC at the DOE Lawrence Berkeley National Laboratory. The machine is based on the HPE Cray “Shasta” platform and a heterogeneous system with both GPU-accelerated and CPU-only nodes. Among the trends noticed by the list compilers, a significant drop in the number of Chinese systems from last edition, and a marked increase in the use of AMD processors. AMD itself provided details about this latter aspect, noting that EPYC processors power nearly 5x more systems compared to the June 2020 list, and more than double the number of systems compared to November 2020. As well, AMD EPYC processors power half of the 58 new entries on the June 2021 list. Intel, also, used the 2021 International Supercomputing Conference for some HPC-related announcements; among them, the next-generation of Intel Xeon Scalable processors (code-named “Sapphire Rapids”) will offer integrated High Bandwidth Memory.

New MLPerf training results

MLCommons, the consortium behind MLPerf machine learning benchmark suite, has released new results for MLPerf Training v1.0. MLPerf Training measures the time it takes to train machine learning models to a standard quality target in a variety of tasks including image classification, object detection, NLP, recommendation, and reinforcement learning. In its fourth round, MLCommons added two new benchmarks to evaluate the performance of speech-to-text and 3D medical imaging tasks. Submissions this round included software and hardware innovations from Dell, Fujitsu, Gigabyte, Google, Graphcore, Habana Labs, Inspur, Intel, Lenovo, Nettrix, NVIDIA, PCL & PKU (Peng Cheng Laboratory & Peking University), and Supermicro. Here are the results, and here is an analysis from EETimes.

Acquisitions

Illinois-based Gamma Technologies, a vendor of multi-physics system simulation software, has acquired Power Design Technologies of Toulouse, France, a specialist in engineering software for power electronics. PDT is also the maker of PowerForge, a SaaS solution for power converter design.

Nexperia, a subsidiary of Chinese giant Wingtech Technology, has completed the transaction to acquire UK-based Newport Wafer Fab (NWF). Nexperia is a customer of the foundry services offered by Newport Wafer Fab and became its second largest shareholder in 2019. The Newport site complements Nexperia’s other European manufacturing operations in Manchester and Hamburg. Originally named INMOS, NWF has a current capacity of over 35,000 200 mm wafer starts per month covering technologies ranging from MOSFETs and Trench IGBTs using wafer thinning methods to CMOS, analogue and compound semiconductors. This new acquisition, however, is reportedly under scrutiny from the UK government.

Texas Instruments has signed an agreement to acquire Micron Technology’s 300-mm fab in Lehi, Utah, for $900 million. This will be TI’s fourth 300-mm fab. It will start with 65-nm and 45-nm production for TI’s analog and embedded processing products and will be able to go beyond those nodes as required.

The world’s 728 unicorn companies in one infographic

Lastly, let’s take a look at the infographic compiled by market research firm CB Insights that visualizes the world’s 728 private companies with billion-dollar valuations. The map sorts unicorns into 15 categories, from auto & transportation to supply chain & logistics. The “Other” category includes renewable energy, space tech, ad tech, and more. Graphics designers, too, will find this infographic interesting as a collection of new company logos.

Credit: CB Insights

This entry was posted on Friday, July 9th, 2021 at 9:57 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.