    Get the competitive edge for AI, data center, business computing solutions & gaming with AMD processors, graphics, FPGAs, Adaptive SOCs, & software.

The world’s most advanced processors
For the world’s most ambitious ideas

AMD processors power people to lead their fields at the cutting edge. From healthcare and entertainment to science and autonomous driving—when AMD performance meets the potential of our partners, we can solve the world’s most important challenges.

together we advance_

Chris Gandolfo, EVP of OCI and AI Sales at Oracle, and Mark Papermaster explore what it really takes to train and deploy large language models at scale. From evolving compute needs and energy efficiency to the rise of inference, this episode dives deep into the future of enterprise AI infrastructure.

00:00 Series Intro & Host Introduction 
00:12 Meet Chris Gandolfo – EVP at Oracle 
01:19 AI at an Inflection Point – Oracle’s Perspective 
04:48 Does Training Ever End? 
05:27 OCI’s Late Entry Strategy – Learning from Rivals 
07:24 Making it Easy for Enterprise 
09:17 Operating in a Scarce Environment 
11:40 How Far Along is Enterprise AI Adoption? 
14:37 It’s a Really Great Time to be a Customer 
15:03 AMD + Oracle: Performance-Driven Partnership 
17:53 Cross Collaboration Across the Ecosystem is King 
20:27 Enabling Edge Inference with Fewer GPUs 
21:59 Co-Innovation on MI355 and Future Roadmaps 
24:08 Openness: Freedom from Lock-In 
25:07 The Future of AI Training and Inference 
26:37 Societal Impact: Guardrails &Responsibility 
28:15 Final Reflections and Episode Close 

***

Subscribe: https://bit.ly/Subscribe_to_AMD 
Join the AMD Red Team Discord Server:   / discord  
Like us on Facebook: https://bit.ly/AMD_on_Facebook
Follow us on Twitter: https://bit.ly/AMD_On_Twitter 
Follow us on Twitch:   / amd   
Follow us on LinkedIn: https://bit.ly/AMD_on_Linkedin 
Follow us on Instagram: https://bit.ly/AMD_on_Instagram

©2025 Advanced Micro Devices, Inc. AMD, the AMD Arrow Logo, and combinations thereof are trademarks of Advanced Micro Devices, Inc. in the United States and other jurisdictions. Other names are for informational purposes only and may be trademarks of their respective owners.

Advanced Insights S2E4: Deploying Intelligence at Scale

From the intricacies of designing an exascale supercomputing system to exploring the workloads these powerhouses excel in, Bridget and Cavin dig into what it takes to create the most super supercomputer of all on the heels of the November 2024 Top500 announcement of El Capitan as the world’s most powerful supercomputer. El Capitan is comprised of over 10 million AMD EPYC™ and AMD Instinct™ cores for a total of over 2.7 Exaflops of raw computing power. Learn more at https://www.amd.com/en/solutions/high...

02:32 What is FLOPS and why does it matter 
03:56 Supercomputing vs. traditional computing in high-stakes workloads
04:53 Our interview with Supercomputing expert, Nick Malaya
05:40 Designing a Supercomputer
07:20 Challenges for the next era of computing

Creating the most super Supercomputer | The Bring Up

We Bring Up: what differentiates a NPU from its CPU and GPU counterparts and how having an AI-enabled laptop can enhance your tabletop gaming experience (amongst other more practical applications). 

00:00:00 Introduction
00:00:53 What is a NPU (Neural Processing Unit)? 
00:03:18 Demo Tabletop Gaming enhanced by an AMD Ryzen™ AI 300 Series Laptop 
00:05:50 How the demo works: Whisper, LLM and Stable Diffusion
00:08:54 AI and accessibility

The Bring Up: Elevating tabletop gaming with a NPU

We Bring Up: Cavin NUKES A CAMERA, and we learn about the exciting process of radiation-hardening spacefaring electronics. Pierre Maillard joins Cavin in his travels to Crocker Nuclear Laboratory in Davis, California to visit the lab where this intense testing occurs. All of this on the exciting conclusion of our two-part episode about SPACE!

The Bring Up: AMD in Space Part 2

We Bring Up: the history and future of AMD's involvement in space travel. Cavin travels to San Jose to visit the labs where the AMD products destined for space are subjected to the extreme temperatures they will experience on their journey to Mars. All of this on the first part of our two-part episode about SPACE!

The Bring Up: AMD in Space Part 1

We Bring Up: a few examples of the incredible technology AMD is enabling in the wonderful world of robotics. Bridget gets the inside scoop at the international FIRST® Robotics Championship where 50,000 high-school roboticists compete for bragging rights. Check out their awesome robot builds that all feature AMD Zynq™. Plus, Cavin builds a robot using the AMD Kria™ SOM that helps detect if he's smiling or not. No, really. He does that.

02:18 FIRST® Robotics competition
06:05 Building a robot with the AMD Kria™ KR260 Robotics Starter Kit
11:08 Seeing the Facial Expression Classification Model in action 

Check out the open source demo here: https://github.com/MakarenaLabs/Facial-Emotion-Classification-Demo

The Bring Up: AMD and Robotics

We Bring Up: insights about AMD Ryzen™ AI, AMD Instinct™ MI300 accelerator and other exciting product announcements from the #AMD #CES2023 opening keynote. From medical treatment to recommending all the things you didn't know you wanted, our CEO Dr. Lisa Su gets to the core of why #AI is all the rage in tech right now. Plus, Cavin breaks down his new flight case PC build shenanigans. What would you name it? Cavin asked ChatGPT and the results were not too shabby.

Discover more: https://www.amd.com/en/events/ces

Chapters
01:13 Our Interview with Dr. Lisa Su
01:24 25 billion with a B transistors and Moore's Law is not dead
02:20 Ryzen™ AI and the role of AI in consumer products
03:50 What’s in store for 2023: Integrating AMD Xilinx IP across the AMD technology portfolio  
04:50 Big thing, very big thing: AMD Instinct™ MI300 accelerator with 3D stacked chiplets!
06:31 ChatGPT is just the beginning
07:03 The low-down on AMD Ryzen™ 7000 X3D
09:36 Cyberdeck PC build break-down by Cavin
12:40 ChatGPT tries to name the Cyberdeck

The Bring Up: Dr. Lisa Su and Cavin's DIY Flight Case PC Build

Sean McClain, Founder and CEO of Absci, and Mark Papermaster step into the world of drug discovery, where processes are incredibly time consuming, costly and complex.

00:00-00:13 Series Overview 
00:14-00:20 On This Episode Description 
00:21-00:55 Intro to Guest Sean McClain
00:56-03:48 The Drug Discovery Process Explained 
03:49-06:05 Impetus for Absci: Scaling Drug Functionality 
06:06-06:39 Complex Drug Permutation Simplified 
06:40-08:26 AI Accelerates the Wet Lab Loop 
08:27-13:54 Groundbreaking Biology Enabled by AI 
13:55-16:59 High Memory is Critical in Model Training 
17:00-17:38 Accurate Models Improve Drug Discovery Success 
17:39-18:35 Lower Cost and Time Investments Required 
18:36-20:15 How Does Technology Disrupt Biological Research? 
20:16–21:31 How Models are Trained with Various Data  
21:32-23:42 Enhancing Compute Efficiency Boosts Scalability 
23:43–25:20 Ten-Year Vision for AI & Biology Intersection 
25:21–26:29 A Future of Personalized Drug Discovery 
26:30–27:17 Key Takeaways  
27:18–27:59 Series Closing

Advanced Insights S2E3: Absci on AI Accelerated Drug Discovery

Dan Rabinovitsj, VP of Infrastructure at Meta, and Mark Papermaster explore the amount of processing power it takes to enable the 3.2 billion people who connect daily on Facebook, Instagram and WhatsApp.

Advanced Insights S2E2: Meta on

Brad McCredie, SVP of AMD’s Data Center GPU and Accelerated Processing business, and Mark Papermaster delve into the world of Exascale supercomputing and AMD’s journey to powering the top two supercomputers in the world. They also discuss the future of high-performance computing and AI.

00:00 - 00:17 Series Overview 
00:18 - 00:33 On This Episode Description 
00:34 - 02:11 Intro to Guest Dr. Bradley McCredie 
02:12 - 02:58 What it Takes to Build a Supercomputer 
02:59 - 05:36 Early Lessons Learned at IBM 
05:37 - 07:22 AMD Journey to Exascale 
07:23 - 08:27 Frontier Supercomputer at Oakridge National Laboratory 
08:28 - 09:12 Frontier Achieves 1.1 Exaflops…and Beyond 
09:13 - 10:15 El Capitan at Lawrence Livermore National Labs 
10:16 - 11:40 AI and HPC Technical Convergence 
11:41 - 13:04 How Many Chiplets Does it Take? 
13:05 - 13:34 Shared Memory Benefits 
13:35 - 15:15 One Size Does Not Fit All, Flexibility for AI 
15:16 - 16:53 Delivering Efficiency with Chiplets 
16:54 - 19:33 Separate Machines for HPC and AI? 
19:34 - 23:16 Driving the Future with Scale and Efficiency 
23:17 - 23:42 Recapping the Main Takeaways 
23:43 - 24:24 Series Closing

Advanced Insights S2E1: AMD on Exascale Computing

Ely Greenfield, CTO of Adobe’s Digital Media business, and Mark Papermaster explore how AI-enabled software and PCs can take on time consuming routine tasks, freeing up creators to work anywhere and focus more on the tasks they really enjoy. They also delve into the critical topics of AI generated art and protecting creatives' original works.

00:00 to 00:16 Series Overview
00:17 to 00:37 On This Episode Description
00:38 to 01:49 Intro to Guest Ely Greenfield
01:50 to 3:34 An Old Story in New, Exciting Time
03:35 to 8:54 Breaking Through Generative AI Barriers
8:55 to 11:35 What AI Changes Mean for Creatives
11:36 to 12:50 Best AI is the One Available in Your Workflow
12:51 to 15:19 Creatives Can Iterate or Churn Out Variations Fast
15:20 to 18:00 Not Everyone Likes AI, Job Concerns
18:02 to 21:22 Boosting PCs with AI Acceleration
21:23 to 22:00 Edge, Local and Cloud Computing
22:01 to 25:56 Industry Collaboration and Open Approaches
25:57 to 26:27 Recapping Main Takeaways
26:28 to 27:01 Series Closing

Advanced Insights Ep. 8: Adobe on Freeing Up Creatives with AI

Mark Russinovich, CTO at Microsoft Azure, and Mark Papermaster discuss how the industry is creating an ecosystem where hardware, firmware and software can be encrypted to expand the chain of trust. They also explored how AI will drive the rapid adoption of these security advances.

00:00 to 00:16 Series Overview 
00:17 to 00:30 On This Episode Description 
00:31 to 00:43  Series Graphic and Music 
00:45 to 01:22 Intro to Guest Mark Russinovich  
01:23 to 02:59 Mark R’s Cyber Security Thrillers  
03:00 to 03:18 Art Imitates Life  
03:19 to 06:30 Protecting Vital Assets 
06:31 to 08:41 Trusted Execution Environment  
08:42 to 10:58 Project Caliptra Industry Standard  
10:59 to 13:19 Collaborative Competition  
13:20 to 14:52 Computing with Confidence  
14:53 to 16:26 Measurement and Integrity  
16:27 to 18:46 Killer App for Confidential Computing   
18:47 to 19:33 Watch What Touches Data Where 
19:34 to 21:09 Protecting AI Models 
21:10 to 22:06 Confidential Multi-party Computations 
22:07 to 22:53 ISV Confidential Clean Rooms  
22:54 to 25:24 Confidential Computing Client to Edge   
25:25 to 27:17 The Next 5-10 Years 
27:18 to 29:05 Thanks Mark and Main Takeaways

Advanced Insights Ep. 7: Microsoft Azure on Confidential Computing

John Roese, Global Chief Technology Officer and Chief AI Officer at Dell Technologies and Mark Papermaster delve into how AI is driving modern system design to tackle the incredible demands of AI computation in the cloud, at the edge and in end user devices.

00:00 to 00:16 Series Overview 
00:17 to 00:30  On This Episode Description 
00:31 to 00:44  Series Graphic and Music 
00:45 to 01:22 Intro to Guest John Roese 
01:23 to 02:13 Disruptions of the Past 
01:24 to 03:20 AI Biggest Inflection Ever 
03:21 to 04:06 Moving Work to Machines 
04:07 to 05:08 Incremental Gains Until ChatGPT 
05:09 to 06:07 John’s Most Interesting Job 
06:08 to 09:23 A Pattern Among Start Ups 
09:24 to 10:22 Assembling the Right Team 
10:23 to 12:13 Working with Agents 
12:14 to 13:59 Responsible AI  
14:00 to 14:47 High Barrier for Enterprises 
14:48 to 16:03 What Makes You Special? 
16:04 to 17:59 How to Deploy AI Successfully  
18:00 to 20:58 Computer System Design Evolves  
20:59 to 22:32 A New Class of Workloads  
22:33 to 27:16 Next Five Years for Enterprises 
27:17 to 28:25 Thanks John and Main Takeaways  
28:26 to 28:54 Closing

Advanced Insights Ep. 6: Dell on AI and System Design

Deirdre Hanford, CEO and Trustee of Natcast, and Mark Papermaster discuss their work on the IAC committee, the importance of public-private partnership and Natcast’s plans to advance research, develop the workforce and keep the U.S. at the forefront of semiconductor innovation.

00:00 - 01:10 - Welcome & Deirdre Bio 
01:11 - 02:07 - Shared Experiences 
02:08 - 02:31 - How they Met 
02:32 - 03:13 - Deirdre’s Start in Semis  
03:14 - 03:41 - People and Problem Solving  
03:42 - 04:21 - Joint Department of Commerce Work 
04:22 - 06:44 - Public Private Partnership Group 
06:45 - 08:32 - Leading Natcast 
08:33 - 09:00 - $11B Opportunity 
09:01 - 09:24 - Deciding What to Fund 
09:25 - 11:14 - Workforce, Workforce, Workforce  
11:15 - 12:34 - Talent is Job #1 at AMD 
12:35 - 14:13 - Helping Chip Designers Get Started 
14:14 - 15:41 - Establishing a Research Agenda 
15:42 - 16:11 - Bridging the Gap Between Lab to Fab 
16:12 - 17:09 - Thought Provoking Conversation  
17:10 - 19:19 - Uniting Industry on a Common Goal 
19:20 - 20:19 - Security and Standards 
20:20 - 20:56 - Data Center Challenges

Advanced Insights Ep. 5: Natcast on U.S. Semiconductor Innovation

Amin Vahdat, Engineering Fellow and VP of Machine Learning, Systems and Cloud AI at Google, and I delve into the transformative era of distributed computing. We explore the key characteristics, challenges and opportunities, as well as the future landscape.

0:00 – 0:30  Introduction
0:30 – 1:47  Amin’s Bio
1:48 – 2:53  Shared experience
2:53 – 3:34  Leapfrogging Innovation
3:35 – 4:59  1970 Blank Canvas
4:59 – 5:52  Innovation Impact and challenges   
5:52 – 6:58  Supercomputing and Connecting People 
6:58 – 7:33  Moore’s Law and the Personal Computing Era
7:33 – 7:47  Connected Computers
7:47 – 8:31  More Powerful PCs
8:31 – 9:07  2000s and Doubling Slows
9:07 – 11:14  Scale Out Takes Hold and Cluster-based Computing
11:14 – 12:28 Totality of Human Knowledge
12:28 – 13:37 Demand for More Computing
13:37 – 14:22 New Types of Accelerators
14:22 – 15:17 30 Seconds/Day = 10 Googles
15:17 – 15:50 5th Generation TPUs
15:50 – 16:17 Transformers Organizing Information
16:17 – 16:58 Model Growth 10X Annually
16:58 – 19:45 Specialized Accelerators for the Win?
19:45 – 20:55 Flavors of General Purpose 
20:55 – 22:36 Power Challenge and Google’s 2030 power goals
22:36 – 25:13 Power Provisioning
25:13 – 26:52 Open Source vs Proprietary
26:52 – 27:58 Societal Infrastructure
27:58 – 29:21 Importance of Team Culture
29:22 – 30:30 Growth Mindset vs Static Thinking 
30:31 – 31:55 Dangers of Village Mentality 
31:56 – 32:34 Communicating Culture
32:35 – 33:45 Thank You Amin
33:46 – 34:29 Episode Wrap Up

Advanced Insights Ep. 4: Google on 5th Epoch of Distributed Computing

On this episode, Dr. Y.J. Mii, EVP and Co-COO of TSMC, and Mark Papermaster talk about what’s beyond 2nm. Have we run into the fundamental limits of physics or will innovation prevail again?

Advanced Insights Ep. 3: TSMC on 2nm and Beyond

Get the most out of your investment in GPUs! 

AMD Instinct™ MI300X (AMD EPYC™ 9575F as host node CPU)  
vs.  
AMD Instinct™ MI300X (Intel® Xeon® Platinum 8592+ as host node CPU)

RESULTS: 
~1.20X the AI Training throughput performance on SDXL v2 
~1.08X the AI Inference performance on Llama 3.1-70B 

https://www.amd.com/content/dam/amd/en/documents/epyc-business-docs/white-papers/maximize-ai-gpu-efficiency-with-amd-epyc-processors.pdf

AI Training & AI Inference performance - Does the CPU matter?

AMD EPYC™ 9965 vs Intel® Xeon® Platinum 8592+ 
Achieve ~1.94X the Generative AI throughput performance (TOKENS/SECOND) with AMD EPYC. 

2025 WHITEPAPER: MAXIMIZE GPU EFFICIENCY WITH AMD EPYC™ HIGH-FREQUENCY PROCESSORS 
https://www.amd.com/content/dam/amd/en/documents/epyc-business-docs/white-papers/maximize-ai-gpu-efficiency-with-amd-epyc-processors.pdf

AMD EPYC™ 9965 running Llama 3.1-8B (BF16, Batch size 4)

To build long term productivity and growth, ecosystems benefit from an Open, Proven, and Ready software stack. 

AMD Developer Central & Tools - https://www.amd.com/en/developer.html
AMD ROCm™ Developer Hub - https://www.amd.com/en/developer/resources/rocm-hub.html
AMD ROCm™ Software - https://www.amd.com/en/products/software/rocm.html
AMD Infinity Hub (Models, PyTorch, Llama, DeepSeek, & more) - https://www.amd.com/en/developer/resources/infinity-hub.html
AMD Instinct Data Center GPU Documentation - https://instinct.docs.amd.com/latest/

Get more out of your parallel computing investment and find out 'why' AMD - Robert Robey, Global Training Lead

Developers

AppDev

DevSecOps

Modernization

Benchmarking

The application development community features top thought leadership focusing on optimal practices in software development, SDLC methodology, mobile app development and application development platforms and tools. Join top software engineers and coders as they cover emerging trends in everything from enterprise app development to developing for mobile platforms such as Android and iOS.

Application Development

As an IT professional, many of the problems you face are multifaceted, complex and don’t lend themselves to simple solutions. The information technology community features useful and free information technology resources. Join to browse thousands of videos and webinars on ITIL best practices, IT security strategy and more presented by leading CTOs, CIOs and other technology experts.

Get more out of your parallel computing investment and find out 'why' AMD - Robert Robey, Global Training Lead

Presented by

About this talk

AMD