Subscribe
Sign in
Home
Accelerator Industry Model
AI Cloud TCO Model
Datacenter Industry Model
Wafer Fab Model
Compliance Policies
Archive
About
Latest
Top
Google Gemini Eats The World – Gemini Smashes GPT-4 By 5X, The GPU-Poors
Compute Resources That Make Everyone Look GPU-Poor
Aug 28, 2023
•
Dylan Patel
and
Daniel Nishball
28
How Nvidia’s CUDA Monopoly In Machine Learning Is Breaking - OpenAI Triton And PyTorch 2.0
Over the last decade, the landscape of machine learning software development has undergone significant changes. Many frameworks have come and gone, but…
Jan 16, 2023
•
Dylan Patel
30
GPT-4 Architecture, Infrastructure, Training Dataset, Costs, Vision, MoE
Demystifying GPT-4: The engineering tradeoffs that led OpenAI to their architecture.
Jul 10, 2023
•
Dylan Patel
and
Gerald Wong
50
Nvidia's Blackwell Reworked - Shipment Delays & GB200A Reworked Platforms
MGX GB200A NVL36, B102, B20, CoWoS-L, CoWoS-S, GB200A NVL64, ConnectX-8, Liquid Cooling vs Air Cooling, NVLink Backplane, PCB, CCL, Substrate, BMC…
Aug 4, 2024
•
Dylan Patel
,
Wega Chu
,
Daniel Nishball
,
Myron Xie
, and
Chaolien Tseng
13
Multi-Datacenter Training: OpenAI's Ambitious Plan To Beat Google's Infrastructure
Gigawatt Clusters, Telecom Networking, Long Haul Fiber, Hierarchical & Asynchronous SGD, Distributed Infrastructure WinnersGigawatt Clusters, Telecom…
Sep 4, 2024
•
Dylan Patel
,
Daniel Nishball
, and
Jeremie Eliahou Ontiveros
17
100,000 H100 Clusters: Power, Network Topology, Ethernet vs InfiniBand, Reliability, Failures, Checkpointing
Frontier Model Scaling Challenges and Requirements, Fault Recovery through Memory Reconstruction, Rack Layouts
Jun 17, 2024
•
Dylan Patel
and
Daniel Nishball
24
AI Datacenter Energy Dilemma - Race for AI Datacenter Space
Gigawatt Dreams and Matroyshka Brains Limited By Datacenters Not Chips
Mar 13, 2024
•
Dylan Patel
,
Daniel Nishball
, and
Jeremie Eliahou Ontiveros
31
GB200 Hardware Architecture - Component Supply Chain & BOM
Hyperscale customization, NVLink Backplane, NVL36, NVL72, NVL576, PCIe Retimers, Switches, Optics, DSP, PCB, InfiniBand/Ethernet, Substrate, CCL, CDU…
Jul 17, 2024
•
Dylan Patel
,
Wega Chu
,
Chaolien Tseng
,
Myron Xie
,
Jeremie Eliahou Ontiveros
, and
Daniel Nishball
22
Datacenter Anatomy Part 1: Electrical Systems
Meta Datacenter Scrapped, Vertiv, Schneider Electric, Eaton, Datacenter Bill Of Materials By Component, Transformers, Switchgear, Redundancy, UPS, ATS…
Oct 14, 2024
•
Dylan Patel
,
Jeremie Eliahou Ontiveros
, and
Daniel Nishball
8
The Memory Wall: Past, Present, and Future of DRAM
Winners & Losers in the 3D DRAM Revolution
Sep 3, 2024
•
Dylan Patel
,
Jeff Koch
,
Tanj
,
Wega Chu
, and
Afzal Ahmad
6
AI Neocloud Playbook and Anatomy
H100 Rental Price Cuts, AI Neocloud Giants and Emerging Neoclouds, H100 Cluster Bill of Materials and Cluster Deployment, Day to Day Operations, Cost…
Oct 3, 2024
•
Dylan Patel
and
Daniel Nishball
10
Nvidia’s Plans To Crush Competition – B100, “X100”, H200, 224G SerDes, OCS, CPO, PCIe 7.0, HBM3E
Roadmap, Supply, Anti-competitive: AMD, Broadcom, Google, Amazon, and Microsoft Have Their Work Cutout For Them
Oct 10, 2023
•
Dylan Patel
and
Myron Xie
21
Share
Copy link
Facebook
Email
Notes
More
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts