Let’s finish the year with a quick recap of some of 2023 events and trends.
New fab announcements
Geopolitical tensions continued throughout the year, prompting governments around the world to pass new subsidies – after the ones approved in 2022 – to support their respective semiconductor industries. Among them, the European Commission launched an €8.1 billion “Important Project of Common European Interest” on microelectronics; Japan is expected to invest $13 billion on semiconductors; and South Korea passed a “K Chips Act” tax credit plan.
The subsidies launched in 2022 – such as the US Chips Act and its European counterpart – spurred several new fab announcements in 2023: a partial list includes a Wolfspeed wafer fab in Saarland, Germany; a Texas Instruments wafer fab in Lehi, Utah; the expansion of an existing Microchip facility in Colorado Springs; a new capital injection to TSMC Arizona; an Infineon plant for analog/mixed-signal technologies and power semiconductors in Dresden, Germany; the choice of Chitose, Japan, as the site for the first Rapidus fab; the expansion of Applied Materials’ campus in Silicon Valley to build the “Equipment and Process Innovation and Commercialization” Center; the expansion of Analog Devices’ European headquarters in Limerick, Ireland, to build a new R&D and manufacturing facility; a joint STMicroelectronics-GlobalFoundries fab in Crolles, France; Intel factories in Kiryat Gat, Israel, and in Wrocław, Poland; Broadcom investments in Spain to build back-end semiconductors facilities; a CEA-Leti FD-SOI fab in Grenoble; the expansion of Analog Devices’ wafer fab in Beaverton, Oregon; a Silicon Box packaging facility in Singapore; the expansion of Infineon’s Kulim, Malaysia, SiC fab; a joint investment from TSMC, Bosch, Infineon and NXP on “European Semiconductor Manufacturing Company”, to be based in Dresden, Germany; an Amkor packaging and test facility in Peoria, Arizona. This is a list of announcements; real fabs will obviously take years to complete.
Artificial intelligence takes center stage this week, with announcements from AMD, some edge AI news and the introduction of Gemini. Among other updates, the recent funding increase to private semiconductor firms.
AMD MI300 to challenge Nvidia H100
At the recent “Advancing AI” event, AMD has launched multiple new AI products, including the Instinct MI300 Series data center AI accelerators, ROCm 6 open software stack, and Ryzen 8040 Series processors with Ryzen AI. The event was intended to showcase growing momentum for AMD-powered AI solutions, particularly in data centers; on the occasion Microsoft, Dell Technologies, HPE, Lenovo, Meta, Oracle, Supermicro and others announced their adoption of the new AMD Instinct MI300X and MI300A data center AI accelerators for training and inference solutions. Analysis website SemiAnalysis has provided a benchmark-by-benchmark comparison of AMD MI300 performance versus Nvidia H100 GPU, highlighting that some AMD results only apply to the “forward pass” processing step. However, in terms of performance – including the software stack, a key element to take advantage of raw hardware power – SemiAnalysis sees AMD “rapidly improving.” SemiAnalysis also highlighted that OpenAI plans to support AMD’s GPUs, including MI300, in the standard Triton distribution starting with the upcoming 3.0 release.
AMD joins the “AI PC” race
Besides Intel, AMD is also touting the “AI PC” concept – that is, adding AI capabilities to mainstream x86 PC processors. The new AMD Ryzen 8040 Series mobile processors feature an integrated Ryzen AI NPU on-die on select models, offering up to 1.6x more AI processing performance than prior AMD models. AMD is also making Ryzen AI3 Software available for users to build and deploy machine learning models on their AI PCs. Laptops with Ryzen AI can offload AI models to the NPU, thereby freeing up the CPU to reduce power consumption while extending battery life. AMD Ryzen 8040 Series processors are expected to be available from OEMs including Acer, Asus, Dell, HP, Lenovo, and Razer, beginning in Q1 2024.
Let’s start this week’s news roundup with a “chip war” update which does not concern China. The Biden administration has reportedly forced a Saudi venture capital firm to sell its shares in Silicon Valley AI chip startup Rain Neuromorphics. According to Reuters, the move follows other actions the US has taken to slow AI development in the Middle East.
Smartphone market is set for an upturn
According to International Data Corporation (IDC), the drop in smartphone sales is finally over: the market research firm expects worldwide smartphone shipments to grow 7.3% year-over-year in the fourth quarter of 2023. The market recovery will continue in 2024 with 3.8% growth expected, followed by low single-digit growth for the rest of the forecast period, resulting in a five-year compound annual growth rate of 1.4%. According to IDC, the smartphone sector is entering the new era of low single-digit growth and lengthened refresh cycles, as the market is maturing. While the total available market will remain below pre-pandemic shipment levels throughout the forecast, average selling prices and market value will remain significantly higher than before.
Credit: IDC
Synopsys 2023 revenues
Synopsys has reported results for its fourth quarter and fiscal year 2023. Revenue for the fourth quarter of fiscal year 2023 was $1.599 billion, compared to $1.284 billion for the fourth quarter of fiscal year 2022. Revenue for fiscal year 2023 was $5.843 billion, an increase of approximately 15% from $5.082 billion in fiscal year 2022.
Will artificial intelligence ever be able to replicate biological brains? It’s a fact that research is advancing on all fronts. On the one hand, the semiconductor industry continues to push transistor size reduction – examples can also be found in some of this year’s IEDM papers – paving the way to systems comprising an ever larger number of transistors. On the other hand, scientists have started bridging artificial intelligence and neurosciences to explore overall network structures – as in recent research from Cambridge University – extending the concept of “neuromorphic” beyond using spiking neurons. This week we will briefly touch these topics – but first, a couple of EDA-related updates.
Cloud-based simulation of mechanical stress for TSMC’s 3D packaging
Ansys has collaborated with TSMC and Microsoft to validate a joint solution for analyzing mechanical stresses in multi-die 3D-IC systems manufactured with TSMC’s 3DFabric advanced packaging technologies. The solution is based on Ansys Mechanical finite element analysis software running on Microsoft Azure cloud infrastructure. 3D-IC systems often have large temperature gradients that lead to intense mechanical stresses between components due to differential thermal expansion. These stresses can lead to cracking or shearing of the connections between various elements. Simulating thermomechanical stress for large and complex devices, while maintaining predictive accuracy, requires substantial computing power.
New release of QuickLogic tool suite
QuickLogic has released version 2.4 of its Aurora eFPGA development tool suite. According to the company, this newest version integrates core tool enhancements that improve the eFPGA utilization and performance of designer’s RTL, particularly in the area of reconfigurable computing. The tool suite integrates fully open-source modules for scalability, longevity, and full code transparency. New features include asymmetric BlockRAM (BRAM) inferencing to reduce the need for manual modification of a user’s RTL design; single stage routing algorithm that boosts maximum operating frequency of a design by up to 24%; power calculation; and other new functionalities.
According to the latest forecast from International Data Corporation (IDC), the semiconductor market has reached a bottom and will soon return to growth. For 2023, the market research firm has revised its September forecast and now estimates that worldwide semiconductor revenue this year will grow to $526.5 billion – down 12% from $598 billion in 2022, but up from the previously estimated $519 billion. For 2024, IDC expects year-over-year growth of 20.2% to $633 billion, up from $626 billion in the prior forecast. Growth in 2024-2026 will be fueled by AI silicon, which by the end of this forecast period will account for almost $200 billion in semiconductor revenues.
Synopsys.ai Copilot
Synopsys has announced Synopsys.ai Copilot, the result of a collaboration with Microsoft to integrate Azure OpenAI Service that brings GenAI into the design process for semiconductors. According to the company, Synopsys.ai Copilot works alongside designers in the Synopsys tools they use every day, enabling conversational intelligence, in natural language, across the design team. Deployable in any on-prem or on-cloud environment, Synopsys.ai Copilot integrates Microsoft Azure on-demand computing infrastructure.
Autodesk’s Fusion is now integrated with Cadence’s Allegro X and OrCAD X
Autodesk’s Fusion mechanical CAD platform has been integrated with Cadence’s Allegro X and OrCAD X PCB design platforms. The collaboration between the two companies aims at solving the problems of current manual design data methods, which require electrical and mechanical engineers to exchange files that can differ from design intent – resulting in errors, unnecessary re-work and costly delays. According to the two companies, the integrated solution enables seamless bi-directional communication between PCB designers and mechanical engineers.
Japan’s government will reportedly allocate roughly $13 billion to boost its semiconductor industry. Part of this money is expected to be used to support TSMC and the recently founded Japanese foundry venture Rapidus. Let’s now move to the other news, which this week includes some announcements from Risc-V Summit North America.
Risc-V updates
Synopsys has extended its ARC Processor IP portfolio to include new Risc-V ARC-V Processor IP. The new Risc-V family includes the 32-bit ARC-V RMX embedded processor IP, scheduled to be available in Q2 of 2024; and the 32-bit ARC-V RHX real-time processor IP and 64-bit ARC-V RPX host processor IP, both scheduled to be available in the second half of 2024. Synopsys also announced it has joined the Risc-V International Board of Directors and Technical Steering Committee.
Ventana has announced the second generation of its Veyron family of datacenter Risc-V processors. According to the company, the new Veyron V2 is the highest performance Risc-V processor available today. It is offered in the form of chiplets and IP. Besides datacenters, V2 targets automotive, 5G, AI, and client applications.
Risc-V starts attracting attention in the context of U.S.-China “chip war”: a bipartisan group of eighteen U.S. lawmakers that includes five Democrats is reportedly asking the Biden administration to prevent China from achieving dominance in Risc-V technology at the expense of U.S. national and economic security. Let’s now move to this week’s news roundup, starting with some EDA updates.
EDA updates: Cadence, Synopsys, Accellera
The new Cadence Voltus InsightAI is – according to the company – the industry’s first generative AI technology that automatically identifies the root cause of EM-IR drop violations early in the design process and selects and implements the most efficient fixes to improve power, performance, and area (PPA). As Cadence maintains, users of Voltus InsightAI can fix up to 95% of violations prior to signoff, leading to a 2X productivity improvement in EM-IR closure.
The new Synopsys Cloud OpenLink program enables chip designers to seamlessly access EDA tools and IP from multiple vendors in the Synopsys Cloud environment. As part of this initiative, the company is releasing an API specification that Synopsys Cloud OpenLink program members can use to deploy system-level integration with a secure and reliable transfer of entitlements to Synopsys Cloud.
Accellera has announced the availability of the Clock Domain Crossing (CDC) Draft Standard 0.1 for public review. This standard aims to ease SOC integration, which often involves combining in-house and externally purchased IPs. The public review is open through December 31, 2023.
Catching up on some of the news from the last twenty-five days or so, this week we report about some interesting EDA related updates – but first, some press reports and company announcements suggesting an upcoming shakeup in the PC processor landscape.
Nvidia to reportedly develop Arm-based processors for Windows PCs
According to a Reuters report, Nvidia has begun designing Arm-based CPUs that would run Microsoft Windows. Nvidia’s initiative is reportedly driven by Microsoft’s effort to help chipmakers build Arm-based processors for Windows PCs, trying to replicate the success that Apple is having with its own Arm-based chips for Mac computers. Qualcomm has been making Arm-based chips for Windows laptops since 2016, but its exclusive license is reportedly expiring in 2024. After that deadline, Microsoft would encourage more chipmakers to join the effort, to avoid relying solely on Qualcomm. AMD is also reportedly planning to make chips for PCs with Arm technology.
Qualcomm’s new PC processor
And Qualcomm has just unveiled the Snapdragon X Elite platform for PCs. It features the custom integrated Qualcomm Oryon CPU and – according to the company – delivers up to two times faster CPU performance versus the competition, matching competitor peak performance with one-third of the power. In terms of AI performance, Qualcomm claims that Snapdragon X Elite can run generative AI models with over 13 billion parameters on-device. PCs powered by Snapdragon X Elite are expected starting mid-2024.
OpenAI, the company behind ChatGPT, is reportedly exploring making its own artificial intelligence chips, possibly through the acquisition of an AI chip company. According to a Reuters report, OpenAI aims to gain independence from expensive Nvidia GPUs.
New EDA releases: Keysight, Nullspace, Mathworks
KeysightEDA 2024 software suite offers three major “shift left” updates. “RF System Explorer” streamlines system and circuit level design workflows for early exploration of system architectures in Advanced Design System; “Digital Pre-Distortion Explorer” and “Digital Pre-Distortion Designer” accelerate wide bandgap power amplifier design and validation using the Dynamic Gain Model; “SystemVue” delivers complete Satcom modeling and simulation solutions for 5G non-terrestrial network, DVB-S2X, and phased array product development.
Nullspace has launched the Nullspace Prep and Nullspace EM 2023.9 release, claiming a 2-4x simulation speed improvement for large problems. The company is also releasing the Nullspace EM Solver on the Windows platform; up until now, the product was only available on Linux. Users interested in the Windows version can apply to take part in a Beta program. Additionally, Nullspace has authored a new whitepaper, “Overcoming Limitations of 3D EM Simulation of Electrically Large Devices.”
Will Silicon Valley’s disruptive innovation capabilities extend to car body manufacturing? In addition to pioneering the use of huge presses with 6,000 to 9,000 tons of clamping pressure, Tesla is reportedly exploring other new solutions to slash the cost of electric vehicles. Technologies being investigated include 3D printing, industrial sand, tailor-made alloys. According to the report, Musk’s goal is to find a way to cast the car’s underbody in one piece.
EDA and IP updates: Zuken, Altair, Ultra Librarian, Intel
Zuken has introduced a three-stage approach to AI-powered PCB design within its CR-8000 platform. The Autonomous Intelligent Place and Route product line introduces a new platform for AI-based place and route, which evolves in stages. “Basic Brain” learns from Zuken’s library of design examples and existing design expertise, and routes the design utilizing the product’s Smart Autorouter based on learned approaches and strategies. In the second stage, Zuken’s “Dynamic Brain” learns from the customer’s PCB designers, utilizing past design examples and integrating them into AI algorithms. The third and final stage is the “Autonomous Brain”, an AI-driven capability that self-improves with each project.
The Ultra Librarian CAD model library is now available to Altair users in several Altair ECAD verification and multiphysics solutions, including PollEx, SimLab, and Altair One UDE. Ultra Librarian gives users instant access to more than 16 million symbols, footprints from a cloud-based library.
And Ultra Librarian has developed a new AI-driven CAD modeling engine to drastically reduce the time it takes to build CAD models.
Intel is launching a new soft processor in the Nios V family targeting its FPGAs: the Nios V/c compact microcontroller – a free, soft-core IP, based on the Risc-V architecture. It will initially target all devices supported in Intel Quartus Prime Pro software with a roadmap to many devices supported in Quartus Prime Standard software.
