AMD CDNA: MI300 & MI400 (Analysis, Speculation and Rumors in 2024)

[Bondrewd]

Couple things looking over the whitepaper again, they doubled the vector (CU) FP32 rate?

Yea.

more VOPD shenanigans?

No it's a more straightforward dual-issue.
940 has no VOPD (the ISA manual is out).

HWS and ACE have known to be firmware running on embedded cores for a few generations, dated back to at least “HWS” was introduced (Polaris?).

Tonga I think.

while a single agent GFX940 configuration can have only 1 HWS by nature

I'm not sure if that's correct given even the monodies have two HWS per die.

 

[trinibwoy]

No it's a more straightforward dual-issue.
940 has no VOPD (the ISA manual is out).

Chipsandcheese delivering the goods as usual with their very detailed CDNA 3 write up. Proper dual issue and an upgraded scheduler with 24!! Wavefronts per SIMD to improve chances of two ready instructions per cycle. As the article states though register bandwidth may be a limitation.

https://chipsandcheese.com/2023/12/17/amds-cdna-3-compute-architecture/

 

[Bondrewd]

Chipsandcheese delivering the goods as usual with their very detailed CDNA 3 write up

I know right.
We're all good friends.

As the article states though register bandwidth may be a limitation.

Same rules as Ampere/Ada/Hopper so yea.

 

[pharma]

Harnessing Incredible AI Compute Power Atop Open-Source Software: 8 x AMD MI300X Accelerators On Linux - Phoronix

www.phoronix.com

 

[Granath]

AMD and Supermicro: Pioneering AI Solutions

In the constantly evolving landscape of AI and machine learning, the synergy between hardware and software is paramount. Enter AMD and Supermicro, two industry titans who have joined forces to empower organizations in the new world of AI with cutting-edge solutions.

www.performance-intensive-computing.com

Unleashing the Open-Source Power of AI through AMD Instinct Accelerators and ROCm

At AMD, we believe that AI is the most important thing to impact computing and technology in the past 50 years, and we remain dedicated to driving innovation and collaboration within the AI industry.   Through our open software ecosystem, ROCm™, and our cutting-edge AI accelerators, we are...

community.amd.com

Multi-node LLM Training on AMD GPUs

Lamini is the enterprise LLM platform for existing software teams to quickly develop and control their own LLMs. Lamini has built-in best practices for specializing LLMs on billions of proprietary documents to improve performance, reduce hallucinations, offer citations, and ensure safety. Lamini...

www.lamini.ai

 

[DavidGraham]

When it becomes not-consistent, they do a reorg and knife a bunch of parts (which, they did! byebye the og mi350x. and byebye the og horrific Si spam MI400).

Looks like MI350 is back on track and is not cancelled.

https://twitter.com/i/web/status/1778436160039100653

 

[Bondrewd]

Looks like MI350 is back on track and is not cancelled.

Not the OG one, no. That one was different XCD n stuff.
Read the poast clearly, not that hard at all!

 

[Newguy]

Hot Chips 2024, Monday 26th August 2024

AMD Instinct MI300X Generative AI Accelerator and Platform Architecture - Alan Smith and Vamsi Krishna Alla, AMD

Advance Program

A Symposium on High Performance Chips

hotchips.org

 

[LordEC911]

MI325X is the HBM3E CDNA3 refresh with 288GB and 6TB/s bandwidth with increased compute coming this year.

MI350 is CDNA4 on 3nm coming in 2025. 35x inference improvement(FP4/FP6) compared to CDNA3.
1.5x more memory and 1.2x "AI compute" TFlops compared to B200.

MI400 is CDNA Next coming in 2026.
AMD Computex 2024 stream link