AWS Silicon Innovation Day 2023
| title | author | length |
|---|---|---|
| Fireside Chat | David Brown, Ruba Borno, Dave Brown | 19m |
AWS describes four custom silicon lines—Nitro, Graviton (Graviton3 in 2022), Trainium and Inferentia (Inferentia2)—and reports over 20 million chips deployed… | ||
| Silicon Development | Nafea Bshara, Gary Szilagyi | 24m |
Annapurna Labs explains how cloud-scale design and manufacturing yield cheaper, more efficient processors: 20 million chips deployed since 2014, custom CPUs… | ||
| Moor Insights Strategy | Patrick Moorhead | 03m |
Analyses Amazon EC2 custom silicon and argues AWS leads public-cloud chips with Graviton Trainium Inferentia Nitro deployed not just in EC2 but across… | ||
| Nitro System Innovation | Anthony Liguori, Ali Saidi | 24m |
AWS describes end-to-end co‑design of Nitro silicon and EC2 software that moves I/O and hypervisor logic onto Nitro cards, freeing host CPUs and enabling 200… | ||
| Fireside Chat | Jeff Barr, Tiffany Wissner | 19m |
Customer examples show concrete wins: a CXM platform moves Amazon OpenSearch and Amazon EKS workloads to Graviton instances, reducing latency 36%, improving… | ||
| Fireside Chat | Raj Pai, Elias Khnaser | 22m |
AWS describes moving virtualization into hardware starting in 2013 and shipping a full Nitro system in 2017 that cuts hypervisor overhead and yields up to ~27%… | ||
| Generative AI Chips | Gadi Hutt, Chetan Kapoor | 24m |
AWS-built chips target inference and training with measured claims: Inf1 delivers ~40% then up to ~70% lower cost versus past EC2 options; Trainium offers 3x… | ||
| Silicon Innovation Day 2023 | AWS On Air ft. AWS Silicon Hugging Face collaboration | Jeff Boudier | 05m |
Collaboration ports Hugging Face libraries to AWS Trainium and Inferentia, using Optimum Neuron and Deep Learning Containers to run real jobs not demos… | ||
| Confidential Compute Architecture | Art Baudo, William Yap, Arvind Raghu | 24m |
AWS describes a purpose-built Nitro architecture that separates host functions into dedicated Nitro cards, a Nitro security chip and a minimal Nitro… | ||
| Customer Spotlight | Art Baudo, Amruta Moktali | 03m |
Skyflow builds a data privacy Vault on AWS Nitro confidential compute using Amazon EC2 Nitro instances to isolate, encrypt and govern sensitive fields (phone… | ||
| Graviton | Martin Yip, Ali Saidi, Stephanie Shyu | 22m |
AWS-built Graviton CPUs use 64-bit Arm Neoverse cores and are running in 19 EC2 instance families across every commercial region; AWS cites >40,000 customers… | ||
| More Than Moore | Ian Cutress | 03m |
Claims that Graviton 3E uses the latest ARM high‑performance cores and targets HPC with 30–35% performance over standard Graviton 3, and about 60% better… | ||
| Trainium and Inferentia | Matthew McClean, Shruti Koparkar | 22m |
AWS Trainium and Inferentia details include TRN1/TRN1N with 512 GB HBM and 13 TB/s HBM bandwidth, Trainium UltraClusters scaling to >30,000 devices and six… | ||
| Inferentia2 | Jason Knight | 02m |
OctoML integrates inf2 instances into its OctoAI service and reports private‑preview deployments that cut latency and cost versus scarce A100 GPU capacity… | ||
| Customer Spotlight | Sairam Menon | 04m |
Johnson & Johnson uses AWS Silicon to run ML that extracts patient experience signals from Janssen therapy data, enabling hundreds of data scientists to train… | ||
| High Performance Computing (HPC) | Art Baudo, Fei Chen, Barry Bolding | 25m |
AWS claims new purpose built chips and instances deliver better price performance and energy efficiency, citing 50 million simulations a week run by Arm and… | ||
| Silicon and Networking | JR Rivers, Madhura Kale | 25m |
AWS uses custom Nitro silicon for both massive active monitoring and for a new transport that reduces tail latency. Nitro cards run 1.5 billion probes per… | ||
| Using Silicon for Cost Optimization | Martin Yip, Pranaya Anshu, Boyd McGeachie | 24m |
AWS claims cloud TCO is 66% lower than on‑prem (63% compute, 69% storage) and highlights concrete wins: Nitro yields ~15% higher throughput; Graviton shows… | ||
| Nitro SSDs and EBS | Art Baudo, Reena Gupta, Prarthana Karmakar | 24m |
AWS describes 60% latency reduction and 75% latency variability drop from custom Nitro SSDs built with a power‑fail‑safe journal inside the flash translation… | ||
| Silicon innovation and the modern network | Peter McKiernan, Brad Casemore | 24m |
AWS claims purpose-built silicon and smart NICs push heavy network work off CPUs, enabling features like congestion control and smart queuing that separate… | ||
| nothing matches | ||