A Call to Arm’s Version 9

EE Times Current · 9 Apr 2021 · 26m

What this episode covers

EE Times editor-in-chief Brian Santo interviews Kevin Krewell, principal analyst at Tirias Research, to dissect Arm's announcement of its Version 9 (V9) processor architecture. Krewell explains that while V9 shares many features with updated V8 versions, it introduces exclusive capabilities like Scalable Vector Extension 2 (SVE2), transactional memory, and Realm Management Platform (RMP) security extensions. These changes signal Arm's strategic push into enterprise-class computing and high-performance data centers, directly competing with Intel Xeon, AMD EPYC, and IBM Power architectures. The discussion highlights how V9 aligns with NVIDIA's acquisition interests by focusing on AI workloads, confidential computing, and performance-per-watt metrics suitable for servers and edge devices.

Mentioned

Companies
Arm, Tirias Research, NVIDIA, AMD, Intel, MediaTek, Apple, Qualcomm, IBM
Products
Ice Lake, Neoverse
Also
Kevin Krewell, SVE2, RMP, RISC-V

Other shows on the same ground, closest first.

  • The Circuit · 41m

    Episode 5: x86 vs. Arm vs. RISC-V, Shifting from General Purpose to Specific Purpose Silicon

    Ben Bajarin and Jay Goldberg discuss the competitive landscape of x86, Arm, and RISC-V architectures, highlighting the industry's shift from general-purpose silicon to specialized, custom designs. They analyze how companies like Ampere Computing, Apple, and Qualcomm are leveraging custom Arm-based architectures to gain performance and efficiency advantages in the data center, challenging Intel's dominance. The hosts also explore RISC-V's rapid adoption, driven by its open nature and cost-effectiveness, particularly among Chinese startups and for IoT and edge AI applications.

  • The Circuit · 39m

    Episode 24: The Arm CPU Datacenter Potential

    Ben Bajarin and Jay Goldberg discuss the potential of Arm CPUs in the data center, highlighting Oracle's decision to port its core database software to Ampere's Arm-based silicon as a major milestone in overcoming software barriers. They analyze the shift toward 'cloud native' workloads, noting that while Arm is competitive with x86 for general web and database applications, the future lies in AI inference and specialized accelerators like Google's VCU. The hosts debate whether this represents an inflection point for Arm's market share, given its licensing model's challenge in extracting value from lower-volume server units compared to mobile devices.

  • The Circuit · 48m

    EP 158: Arm 2.0, Memory Freak Out, and AI Models

    Hosts Ben Bajarin and Jay Goldberg analyze ARM's strategic shift from a pure IP licensor to a merchant silicon vendor with the launch of its AGI CPU, designed for agent-native workloads. They discuss the market implications for competitors like NVIDIA, AMD, and Intel, as well as the reactions of key customers like Meta and OpenAI. The hosts also cover a recent volatility in memory stocks triggered by Google's TurboQuant announcement and speculate on Anthropic's rumored 'Mythos' model and the future of AI ROI.

  • The Circuit · 36m

    Episode 46: A Discussion with Arm CEO Rene Haas

    Rene Haas, CEO of Arm, explains that Arm is the 'brain' powering 70% of the world's electronic devices, operating through a licensing model rather than manufacturing chips. He details a strategic shift from generic CPU designs to productized, purpose-built solutions like Neoverse for data centers and automotive subsystems, which command higher royalties and accelerate time-to-market. Haas highlights growth drivers in cloud AI, PCs, and automotive, while noting that US sanctions on China have limited impact due to Arm's non-US design origin, though RISC-V gains traction in simple IoT applications. The conversation concludes with Arm's close partnership with Intel Foundry Services and the potential for Arm to deliver verified chiplet designs.

  • The Circuit · 30m

    EP 192: Talking Apple Silicon with Apple's Tom Boger, Kaiann Drance, and Sri Santhanam

    Tom Boger, Kaiann Drance, and Sri Santhanam from Apple discuss the transition to 2-nanometer process technology across the A20 Pro and M-series chips, highlighting the shift to a dual neural engine and unified memory architecture. They explain how vertical integration allows Apple to optimize packaging, such as moving DRAM off the SoC to improve thermal management and sustained performance for AI workloads and gaming. The panel also details the implementation of the Secure Enclave and Secure Exclave to handle computationally expensive cybersecurity tasks and protect sensitive data at the silicon level.

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by EE Times Current. The episode itself belongs to its publisher.