AMD Data Center and AI Technology Premiere highlights

On June 13th, 2023, AMD announced the products, strategy and ecosystem partners that will shape the future of computing, highlighting the next phase of data center innovation. AMD was joined on stage with executives from Amazon Web Services (AWS), Citadel, Hugging Face, Meta, Microsoft Azure and PyTorch to showcase the technological partnerships with industry leaders to bring the next generation of high-performance CPU and AI accelerator solutions to market.

***

©2023 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.

[ProjectPhysX]

192GB VRAM on the MI300X. Finally worthy hardware for my CFD simulations! The AI boom is good for one thing after all: creating sufficient demand for more and faster memory.

[anthonybaker]

We are still waiting for FSR 3 and HYPR-RX. Come on AMD.

[KingSquanchy]

Thank you for calling the Fairmont hotel in SF and using our AV service needs. Hope to see you again.

[digibluez]

really, BUT what about FSR 3 and new software that you promoted for the sale of 7000 series GPUS, you told would release at the first half of the year?! where is it?!

[thejustinjustin]

I love these chips. I just got the rog ally and the z1 extreme is pretty amazing.

[johnnm]

Hello AMD please offer a robust set of drivers that are performant and stable. Spend a lot of time and resources on ROCm and similar frameworks. Make it easy for developers to harness the power of your GPUs and APUs

[kaisonix]

This showcase is on the point!!! nVidia showcase can save their cringe showcases to their tiktok but not for the customers/corporate partners

[thepianisthd]

*When Comes HYPER RX Finally Out?? for Radeon Software we waited for so long months for please AMD say something about this announce at least.*

[Sentinar]

When will you finally fix the driver problems for RDNA2 and RDNA3 ? 🥲
When will FSR 3 be released ? 😮‍💨The main competitor is already far ahead.

[ARTEMEX]

Amd молодцы ! После того, как я глубже ознакомился с RDNA 4 моё мнение сменилось об видеокартах RADEON. Давайте AMD дерзайте ! Вы сможете !!!
P.S. пожалуйста создайте больше программного обеспечения в следующем поколение. Надеюсь оно выйдет в конце 2023 или Q1 2024

[MarketAndChurch]

Will this MI300X be enough to win over customers who are buying up Nvidia GPUs to train their Large Language Models?

[matteoromenghi]

Bergamo, up to 128 cores and 82 B transistors. 😲
When a generation of EPYC processors named Parma? 🙂

[duladrop]

What made MI300X a game changing will be it's Raw performance the 5, 218 TFlops is a beast, even the two H100 Hopper on NVLink can't even beat 1 MI300X GPU because it can only get 3K+ TFlops... That is a game changer. AMD have been subjected to a lot of criticism form Nvidia fan boys because they can't make a product that can beat their Nvidia's best products. Well here it is the name is "MI300X Instinct".

[kaisersoza]

When's the 7800xt coming out? Ive been holding onto the cash for ages 😭

[ME-fvug]

For training on a large data corpus, large AI language modules, TensorFlow frameworks are used. Therefore, in graphics accelerators, it is necessary to have hardware, not only a lot of high-speed HBM3 memory, but also the presence of tensor cores, I hope that AMD, in the next generations of its graphics part of accelerators, will create the presence of hardware tensor architecture in their products. But the new-generation AMD Ryzen Threadripper CPUs turned out to be very productive and very energy efficient, yet data center processors are simply indispensable in the current competitive environment in the CPU market.

[yolo]

Please go to all time high! Amd Yes!!!

[netbat]

we need a cpu with few cores with small process technology to focus on achive world record overclocking. a low TDP, low performance but high frequency cpu. just for overclocking record

[InternetListener]

RDNA3 supposedly has twice the TeraFlops od RDNA2 but average gaming goes from 0 to 10% (with +30% maximum) on architecture only comparison, but in computing RDNA3 seems to be half perfromance of CDNA2 to almost twice (it obviously depends on load) showing a great potential, but it seems faster up to twice to RDNA2 doing computing (according to GpuOpen math code)... could you please explain or show where a 7600 is faster at Tflops than a 6800XT? You say so in specifications on your AMD web, but it's not the current experience, at least in games... how RDNA1 or Vega can be as fast as twice Tflops RDNA3 architecture?

[johnnyxiv]

Lisa next time dont say sampling let the market speculates instead.. now the stocks keep go down 😢

[Axyber]

You say all gamers are respectable and you offer your new technology only on your new cards. Are you throwing away a lot of gamers who have rdna 2 cards to sell your new cards? Have you become a liar like Nvidia? I bought amd because of the constant support of old cards This new policy of yours is considered theft from us.