Yuval Bachar

Sierra Forest and Clearwater Forest will be crazy server products, skymont has higher IPC then intels current best single core performance Power Efficiency Gain is solid Compute/Die size will be crazy Power, Performance, Area, Cost only the power metric is unimpressive everything else is crazy AMD Turin cores are less performance at more area Let’s goooo can NVIDIA and Intel please finish off AMD? They are the volvo of semiconductors, nobody likes them #intel #wintel #nvidia #ai #computex

2

1 Comment

New SiFive Performance P870-D Complements Intelligence Chipset for AI TL;DR - SiFive announced the SiFive Performance P870-D, a high-performance RISC-V datacenter processor designed for demanding AI workloads, offering up to 256 cores, enhanced scalability, and advanced Reliability, Availability, and Serviceability (RAS) features. This processor positions SiFive as a competitive player in the data center market, providing data center architects with a flexible, power-efficient solution for AI-driven applications. SiFive aims to leverage RISC-V's open-source flexibility to challenge established industry giants like Intel and AMD. Read the Research Note in the comments below.

1

1 Comment

AI is driving the next wave of growth for the semiconductor industry (and those of us in the industry can't wait to see the anticipated growth 😉). Recently, I was encouraged to watch a keynote featuring Jensen Huang of NVIDIA. While it's highly informative and I recommend watching it in its entirety, what struck me most had nothing to do with AI. Around 36:05, the facilitator asked Jensen to provide advice on how to achieve success. His response is profound, drawing parallels between expectations, resilience, success, intelligence, and character. His commentary on how he leads his 55 direct reports (yes, 55!) is also thought-provoking. Take a few minutes to watch this keynote and listen to Jensen's insights on success and leadership. #artificialintelligence #NVIDIA #leadership

20

2 Comments

I appreciate the opportunity to express my views on behalf of Percepio® on the need to bring Observability Driven Development (ODD) also into the embedded/edge space, the massive heterogenous Operational Technology universe, often "the last mile" also in the most complex IT systems. Thanks to our founder and CTO Johan Kraft for insightful co-authorship, in this article in the New Electronics Outlook supplement, on gracious invitation by editor Neil Tyler. To briefly highlight by quoting myself: "The increasing complexity of embedded systems is challenging traditional development and testing methods ... To address these challenges, the concept of Continuous Observability has emerged. ... ODD requires a “shift left” approach, where observability is designed-in early to support all following stages of product development and maintenance. This is not only a concern for software QA and test management but a key factor in the performance and competitiveness of product development organisations. " (Also a shout out and thanks to Bijit Ghosh for promoting the concept of ODD from a cloud-native perspective). https://lnkd.in/dQrxnecT #odd #observability #embeddedsystems #sdv #iot #edge 

34

5 Comments

Honored to see Dryad Networks featured in Silicon Republic, spotlighting our journey in leveraging IoT technology to combat the escalating threat of wildfires. As the article highlights, wildfires are becoming more frequent and destructive due to climate change, putting communities, ecosystems, and our planet at risk. We’re addressing this crisis with Silvanet, our IoT-based ultra-early wildfire detection system. Unlike traditional methods that often detect fires too late, Silvanet uses a network of gas sensors connected through solar-powered IoT devices to detect wildfires in their earliest stages—before they spread uncontrollably. Key takeaways from the article: - Our technology has the potential to protect forests worldwide, especially in high-risk regions grappling with rising temperatures. - Silvanet’s solar-powered infrastructure and low-power networks make it an accessible solution even in remote areas. - By detecting fires faster, we not only save lives and ecosystems but also reduce CO₂ emissions from forest fires, tackling a major contributor to global climate change. Read more here: https://bit.ly/3ZK1tjj #ClimateTech #IoT #WildfireDetection #Sustainability #Innovation

20

The latest epiosde of our #TantrasMantra #podcast is out... https://lnkd.in/gqiBHvVx In this episode, Leonard Lee of neXt Curve joins me in discussing and analyzing the key takeaways from COMPUTEX TAIPEI and Apple's #WWDC. #Computex Buzz: #CoPilotPlus PCs Take Center Stage: We explore how CoPilot+ PCs overshadowed everything at Computex, show-and-tell of Qualcomm, the announcement of #Lunarlake by Intel Corporation, #Ryzen #AI300 of AMD, both powering X86-based CoPilot+ PCs, opportunities during back-to-school season, changing competitive landscape by Holiday season and more. Apple Intelligence Announcement: We delve into #AppleIntelligence, its current standing against competition, Apple's collaboration with OpenAI, #security and #privacy concerns, benefits of vertical integration in #HybridAI, and more. Arm #WindowsOnArm #GenerativeAI #Genai Tantra Analyst #Ai #AIPC #MeteorLake

11

Growth of Arm in the datacenter starts with availability, then adoption. With some hyperscalers rolling their own designs, everyone else has to look to the open market to see what's for sale. The only real option here is from Ampere, which recently announced a custom 3nm Arm 256-core CPU as part of the AmpereOne family. More cores, more memory channels, and a custom microarchitecture they might disclose later this year. It's a chiplet design too, showcasing where this market is going. https://lnkd.in/e6JMXCum

95

8 Comments

Exciting developments! Discover how Exostellar Inc. is partnering with Amazon Web Services (AWS) to revolutionize semiconductor design, helping organizations significantly cut costs and optimize their processes. With semiconductor design workflows relying on compute-intensive tasks like functional verification, place and route (P&R), physical verification, and electrical signoff, managing costs can be a challenge. Exostellar’s solutions on AWS are designed to provide optimal performance and scalability for cloud-based EDA tools while keeping costs in check. Ready to learn more? Check out this insightful blog to see how we’re making a difference in the industry: https://lnkd.in/g9jTdQWq Nayan Lad Frederic Julien Rohan Prakash Zhiming Shen Nick Cox Tony Shakib Neel Desai #Semiconductor #HPC #CloudComputing #AWS #Exostellar #TechInnovation

17

5G + Edge + GenAI = New Business Model opportunities Meta has released Llama 3.2 lightweight models in 1B and 3B sizes available on #watsonx --> advantages: > Small size is perfect for mobile and edge devices or CPU only inferencing. > Optimized for ARM based processors, support Qualcomm and MediaTek hardware from day one (covering majority of mobile processors out in use). > Enhanced Privacy - No need to send sensitive data such as messages and calendar invite to the cloud, maintaining user privacy. > Supports up to 128K tokens and optimised for summarisation, rewriting, instruction following, language reasoning, and tool use.  > Tool Calling: The 1B and 3B models excel in tool calling, enabling developers to create personalized, on-device applications. > Instantaneous Responses: Local processing makes prompts and responses feel instantaneous. > Tiny size achieved through network pruning and knowledge distillation using larger Llama 3.1 > Post-training step uses supervised fine-tuning (SFT), rejection sampling (RS), and direct preference optimization (DPO) for alignment. Example Use Case: Integrated Applications that could help summarize the last 10 messages received, extract action items, and leverage tool/function calling to directly send calendar invites for follow-up meetings. RIGHT MODEL --> RIGHT USE CASE --> RIGHT GOVERNANCE --> Success !   

25

All in on #Transformer, it is beating all other architectures in end-to-end applications. Like #Etched, #Skymizer is betting on Transformer only accelerator with compiler-centric approach. Skymizer’s first IP product, EdgeThought, is a transformer based on-device LLM inferencing only accelerator.

11

I'm always happy and proud to hear our early investors at REFASHIOND Ventures: The Industrial Transformation Fund talk about the work they are doing. Each of them is having real impact in building supply chains for tomorrow. In this podcast Zac talks about the underlying infrastructure for what what I describe as Information and Data Supply Chains. This is worth listening to if you want to get into some of the issues that will determine the future of Physical Supply Chains. #TWSCF #TNYSCM | The past ran on supply chains. The present runs on supply chains. The future will run on supply chains. The world is a supply chain.®

4

Totally geeked out a couple of weeks back after driving next to a self-driving car 🚖for the first time. Years of watching this tech evolve on visits to the Bay Area, but this hit different! So many engineering marvels at play - it's a win for human ingenuity for sure.🤖 But then the mind wanders... With commutes a thing of the past, what will we do with all this "free" time? More time on Instagram, YouTube ... ? Or a chance to learn more, do more and tackle the big challenges we face ? What are your thoughts on autonomous vehicles and more broadly the future of having more time on our hands ? #futureoftransportation #autonomousdriving #technology #connectedcar https://lnkd.in/gayjuaAB

13

2 Comments

As #AI scales across industries and use cases, understanding how to measure AI performance becomes key in developing solutions with sufficient head room. Like we say, "not all TOPs are created equal". At Qualcomm, we focus on running AI inferencing workloads on our super-efficient and powerful Neural Processing Units (NPUs). Highly recommend reading our latest article on how to measure AI workloads, performance and the right place to run them.  https://lnkd.in/gzV33PCa

75

there simply is not enough large scale interconnect capacity (electrical grid availability, for non energy nerds) to support the near term data center needs of BOTH the AI/LLM arms race AND the non-AI internet. winning AI is of higher strategic importance to big tech & to US geopolitics than the connectivity of a lot of things we use every day — ranging from Netflix to Facebook to Dropbox to Xbox and beyond. Something has to give.

45

1 Comment

Very interesting to note as to the new #HighBandwidthMemory capabilities of the new NVIDIA #Blackwell architecture briefing which allows the new chips to perform at x18 times current CPU and x6 times NVidia H100 capabilities. This is significant for generative AI applications that require high database query demands. Additionally, the architecture also assists with confidential computing by boosting the Trust Execution Environment. #TEE https://lnkd.in/gvGXRmSr

9