HomeBlogAMD CES 2023 Keynote: Power Moves in AI and Processors

AMD CES 2023 Keynote: Power Moves in AI and Processors

Hey tech aficionados! I continue to tune in to the AI tech world straight from the most recent and popular Keynotes. Today is a summary of some game-changing developments in the AMD CES 2023 Keynote.

In this blog post, let’s quickly cover the new core technologies and features of AMD.

AMD Instinct MI300X: The Gen AI Powerhouse

  • MI 300 X takes the stage as the highest performance accelerator for generative AI, riding on the sleek CDNA 3 data center architecture.
  • The magic lies in the details – a new compute engine, sparsity support, and cutting-edge data formats like FP8. The support of sparsity is surprising to me, as for now, Nvidia has been the only one implementing it. 
  • In regards to performance, it has 2.4 times more memory capacity and 1.6 times more memory bandwidth than the competition.

AMD Instinct MI300X Specs Demystified: A Tech Ballet

  • Picture this: Four IO dies in the base layer, 256 MB of Infinity Cache, and all the next-gen IO essentials.
  • Stack eight CDNA 3 accelerator chips on the IO die for a mind-bending 1.3 petaFLOPS of FP16 and 2.6 petaFLOPS of FP8 performance.
  • Then, connect the 304 compute units with dense through silicon vias (TSVs) for up to 17TB/s of bandwidth.
  • And finally, top it off with eight stacks of HBM three, totaling 192 gigabytes of memory at 5.3TB/s of bandwidth.

AMD ROCm 6: Open Source Magic

  • Enter ROCm 6, the modular and open-source hero of the story, waving goodbye to the closed-off vibes of CUDA.
  • The open-source community's speed of light in deploying new algorithms and models shines through, making ROCm 6 a tech marvel.
  • Shipping later this month, ROCm 6 is optimized for generative AI, particularly for LLM, boasting library optimizations, expanded ecosystem support, and performance boosts.

AMD Instinct MI300A: The CPU-GPU Love Story

  • MI300A steps into the spotlight, heralding the fourth-generation infinity architecture.
  • CPU and GPU unite, sharing a unified pool of memory in this Accelerated Processing Unit (APU).
  • The numbers speak volumes – 61 teraFlops of double precision (FP64), 122 teraFlops of single precision (FP32), 128 GB of HBM3 memory, and a bandwidth of 5.3 TB/s.
AMD Instinct MI300A

Hawk Point AMD Ryzen 8040 Series: Mobile Marvels

  • Hawk Point combines all of our industry-leading performance in battery life and it increases AI tops by 60% compared to the previous generation.
  • This 8040 Series will deliver more performance for multi-threaded applications, 1.8 x higher frame rates for games, and 1.4 x faster performance across content creation applications.

Wrap up

All tech giants: Microsoft, Google, Intel, Nvidia, and now, with AMD, we continue witnessing a tech symphony – from MI300X's AI prowess to the open-source magic of ROCm 6 and the CPU-GPU unity in MI300A. 

The future of computing is taking shape, and these innovations are paving the way.


Juan Mulford
Juan Mulford
I have been active in IT for over fourteen years now. I am a solutions architect, working with storage, virtualization, and VDI solutions. For the past ten years, I have been living and working in Taiwan.

Leave a Reply

- Advertisement -

Popular Articles


Raspberry Pi OS in a Virtual Machine with VMware

Although the Raspberry Pi OS is designed and optimized for the Raspberry Pi module, it is possible to test and use it without its hardware, with VMware. This solution can be useful if you are a developer (or just a curious guy) and don't have a Raspberry Pi module with you
Unable to delete inaccessible datastore

Unable to delete an "inaccessible" datastore

I was switching my storage array, so I migrated the VMs from that old datastore/storage to a new datastore/storage. The old datastore was shared by 3 ESXi hosts, no cluster. After migrating the VMs and unmount/delete the datastore, it was still presented in two of the ESXi hosts and was marked as inaccessible.
This is not a valid source path / URL

This is not a valid source path / URL - SourceTree and Gitlab

I have been working on a project with a friend who set up a repository in Gitlab but even though I was able to view all projects on it, I couldn’t really join the repository. I was using SourceTree and Gitlab.

How to export a Virtual Machine using the VMware OVF Tool

The VMware OVF Tool is implemented by VMware for easily importing and exporting virtual machines in Open Virtualization Format (OVF) standard format. Here, I want to show you how to download and install it, and then how to use it from a Windows machine.
Couldn't load private key - Putty key format too new

Couldn't load private key - Putty key format too new

couldn't load private key - Putty key format too new.” This issue happens when you use PuTTygen to generate or convert to a ppk key. Here is how to fix it. 
- Advertisement -

Recent Comments