330

AMD is preparing to make important announcements for the data center market with the Accelerated Data Center Premiere event. There was no information about the content of the virtual event, which will be held on November 8th. However, we expect the company to introduce its next-generation EPYC processor and Instinct server accelerators.

If we start with the processors, there is a good chance that AMD will launch the Milan-X processor family, which will replace the current generation server platforms. The new Milan-X chips will be an extension of the EPYC 7003 product family and will use the company’s highly anticipated 3D V-Cache chip stacking technology. In addition, up to 512 MB of L3 cache can be accommodated, in addition to the maximum 256 MB available with the Milan-X family.

According to previous leaks, the new flagship SKU is the EPYC 7773x with 64-core/128-thread design with 2.2GHz base clock, 3.5GHz boost clock, 280W TDP and 768MB total L3 cache (including 3D-V Cache). it will be.

Expected EPYC Milan-X Processor Features

CPU Core/Imprint Base Clock Boost Clock TDP L3 Cache (L3 + 3D V-Cache)
EPYC 7773X 64/128 2. 2GHz 3. 5GHz 280W 768 MB
EPYC 7573X 32/64 2. 8GHz 3. 6GHz 280W 768 MB
EPYC 7473X 24/48 2. 8GHz 3. 7GHz 240W 768 MB
EPYC 7373X 16/32 3. 05GHz 3. 8GHz 240W 768 MB

Radeon Instinct GPU Family

We have already reported that AMD will launch the Instinct MI200 series based on the CDNA 2 (Aldebaran) architecture when moving to the Instinct family. We’re likely to see products supporting up to 128GB of HBM2E memory with new accelerators.

On the other hand, recent rumors reveal that the company will announce two GPUs, MI250 and MI250X. The MI250X uses TSMC’s advanced 7nm manufacturing technology while boasting 110 CUs per die (2x MCM = 220 CU), 1.7 GHz boost clock and 500W TDP.


Like it? Share with your friends!

330
Michael Lewis

0 Comments

Your email address will not be published. Required fields are marked *