AMD vs. Nvidia: Battle of the AI Chips

Показать описание

We compare the reveals of AMD's MI300X and Nvidia's Grace Hopper AI chips.

Рекомендации по теме

Комментарии

They both look like a brother sister 😅

indraprakash

Skynet will eventually decide which manufacturer to select

MrDebranjandutta

Picture this folks -> MI300X computer power is 5.2K Tflops while Grace hopper has 1.5K Tflops even if two H100 are join via NVLink can get 3.9K still not enough to compete against a Single MI300X. Yes MI300X is truly a beast... I was just even expecting and Hope they can have a 1.5K TFlops GPU so that they can compete with Nvidia, but AMD even created a Powerful Beast that can handle 5K TFlops! Wow. With the ever evolution of ROCm since it's open source this is going to be the start of a good future of AMD in AI space.

duladrop

AMD slides showing Nvidia H100 comparison shows who is the king and who sets gold standard. Apple never compares, Nvidia never compares bcoz these guys set the gold standard while others keep comparing and keep challenging the dominance

gumpenaprasanth

They are not battling. They are cousins. And they are very very close. The chip industry is owned by one family

shinkurt

Same one please write the full spec list for each one

GostFlash

Maybe they're competing bcoz of childhood issues. (Assuming they're family)

blackwolfkills

AMD duron and some of the Memories, upgrades and newer series has this KiCK and JUICE that won't ruin your platform and board, its COOLING POWER is SUPER COOL, since AMD has plenty of juice, while NVIDIA has some juice, yet it caters for specific programs, AMD has plenty of features, its great that they have overcomed some of the issues back then. Infrastructure is going to be as healthy and engineered further for success. Congratulations.

Rod-bpow

Their parents are both from Taiwan and they are relatives . They were both born in Tainan city, Taiwan however his Taiwanese and Chinese is much better than hers.
His son had a bar in Taipei City before.

Brandon.R

Mi300x has 16 matrix cores per SM compare with only 4 matrix core per SM inside Mi250x.

It has supperior 5 PETA OPS INT8 compare with nvidia h100 &h200 only 4 PETA OPS. Also mi300x has higher fp16 and bfloat 16 compare with competition

kusumayogi

Got to heard that Nvidia is collaborating with Mediatek, so what's new is coming?

oliversmith

Lots of people assumed Amd was going to fall off the map, but that is far from the truth.
Amd and Nvidia is operated by the same family they are also owned by the same shareholders.

robcooper