AMD conference reveal: new generation EPYC processor and MI300 graphics card compete with NVIDIA

News on June 9, according to the latest reports, in the global graphics card market, NVIDIA has always been the unshakable overlord. Whether it is game cards or accelerator cards, its market share continues to rise, while the share of competitor AMD constantly being compressed. According to the editor's understanding, AMD is about to launch a new graphics card. This product will bring major changes to the market and cause considerable competitive pressure on Nvidia.

AMD will hold a special press conference on the evening of June 13th (U.S. time)/early morning of June 14th (China time), focusing on announcing its products for data centers. . Among them, what has attracted much attention is the new generation of EPYC processors, which include the 128-core Bergamo and the cache-enhanced Genoa-X processor. The introduction of these processors will bring more powerful computing power and performance to data centers.

In the field of graphics cards, AMD will also reveal the Instinct released last year at the press conference. MI300 graphics card launch date. According to the news, this graphics card is expected to be available in the second half of the year as soon as fall. It is worth mentioning that Instinct The MI300 graphics card will become a major weapon for AMD to compete with NVIDIA, and its performance can compete with NVIDIA's high-performance computing card H100.

It is understood that Nvidia has almost no substantial competitors in the fields of high-performance computing and artificial intelligence. Because popular AI card models such as A100 and H100 are in short supply, their delivery cycle is as long as half a year. The increase in market demand has caused a sharp rise in prices, with the price of H100 rising from 250,000 yuan to a higher level. Many companies have invested a lot of money to compete for these scarce card types to gain an advantage in this hot AI industry.

The MI300 graphics card launched by AMD will become a trump card against Nvidia. The graphics card uses an advanced 5-nanometer manufacturing process and has an astonishing 146 billion transistors. It integrates a CDNA3 architecture GPU unit, 24 Zen4 architecture CPU cores, and large-capacity Infinity Cache unlimited cache, and HBM3 high-bandwidth memory with a capacity of up to 128GB. In addition, it also supports the fourth generation Infinity Fabric bus, CXL 3.0 bus, unified memory architecture and new mathematical calculation format. It is said that the AI performance of the MI300 graphics card is up to 8 times higher than that of the previous generation, which is enough to meet huge computing needs.

With the advent of AMD’s upcoming new generation of EPYC processors and MI300 graphics cards, the global data center and artificial intelligence fields will usher in a new competitive landscape. NVIDIA will face huge competition from AMD, and the market will also usher in more competitors and opportunities. We are eagerly looking forward to the AMD conference and looking forward to unveiling the mystery of these exciting new technology products.

The above is the detailed content of AMD conference reveal: new generation EPYC processor and MI300 graphics card compete with NVIDIA. For more information, please follow other related articles on the PHP Chinese website!