AMD Instinct MI250X with MCM GPU to feature 110 Compute Units, 128GB HBM2e memory, and 500W TDP

Published: 23rd Oct 2021, 18:41 GMT   Comments

Please note that this post is tagged as a rumor.

AMD to launch Instinct MI250X with 48 TFLOPS in FP64

ExecutableFix reveals the first details on MI200 series accelerators from AMD.

According to the leaker, AMD is to launch MI250 and MI250X Instinct accelerators, both based on Aldebaran GPU featuring CDNA2 architecture. The MI250X has been confirmed to feature 110 Compute Units and 128GB of HBM2e memory.

The leaker claims that the accelerator will have a TDP of 500W and will be built using a 7nm process architecture. With 110 Compute Units clocked at 1.7 GHz, the accelerator would offer 47.9 TFLOPs double-precision (FP64) and single-precision (FP32) compute performance and 383 TFLOPS in half-precision calculations (FP16/BF16).

The MI250X and MI200 are supposedly both based on Aldebaran GPU, except the MI250 non-X would have some CUs disabled. It would appear that the configuration of the cut-down part has not yet been confirmed. The MI250X may also be a higher clocked version, which would be a similar approach to NVIDIA SXM variants and their respective PCIe models.

The MI200/250/250 series are to compete with Intel Ponte Vecchio (Xe-HPC) and NVIDIA H100 accelerators, both expected to debut next year.

RUMORED AMD Instinct Accelerators Specifications
Accelerator NameAMD Radeon Instinct MI60AMD Instinct MI100AMD Instinct MI250AMD Instinct MI250XAMD Instinct MI300
Architecture7nm GCN5 (GFX906)7nm CDNA1 (GFX908)7nm CDNA2 (GFX90A)7nm CDNA2 (GFX90A)CDNA3 (?)
CPUZen4 (?)
GPUVega 20ArcturusAldebaran (MCM)Aldebaran (MCM)? (MCM)
Compute Tiles11224
Compute Units64 (64)120< 1101104x (?)
FP32 Cores (Full GPU)4096 (4096)7680 (8192)TBCTBC4x (?)
GPU Clock Speed1800 MHz~1500 MHzTBC~1700 MHzTBC
FP16 Compute29.5 TFLOPS185 TFLOPSTBC383 TFLOPSTBC
FP32 Compute14.7 TFLOPS23.1 TFLOPSTBC47.9 TFLOPSTBC
FP64 Compute7.4 TFLOPS11.5 TFLOPSTBC47.9 TFLOPSTBC
VRAM32 GB HBM232 GB HBM2TBC128 GB HBM2ETBC
Memory Clock1000 MHz1200 MHzTBCTBCTBC
Memory Bus4096-bit4096-bitTBCTBCTBC
Memory Bandwidth1 TB/s1.23 TB/sTBCTBCTBC
Form FactorDual Slot, Full LengthDual Slot, Full LengthTBCOAMTBC
CoolingPassive CoolingPassive CoolingTBCTBCTBC
TDP300W300WTBC500WTBC

Source: ExecutableFix




Comment Policy
  • Comments must be written in English.
  • Comments deemed to be spam or solely promotional in nature will be deleted. Including a link to relevant content is permitted, but comments should be relevant to the post topic.
  • Comments and usernames containing language or concepts that could be deemed offensive will be deleted. Note this may include abusive, threatening, pornographic, offensive, misleading, or libelous language.
  • Comments complaining about the article subject or its source will be removed.
  • A failure to comply with these rules will result in a warning and, in extreme cases, a ban.
  • Please note that comments that attack or harass an individual directly will be deleted and such comments will result in a ban.
  • VideoCardz Moderating Team reserves the right to edit or delete any comments submitted to the site without notice.
  • If you would like to appeal for a comment section ban to be removed, please use this page.
  • If you have any questions about the commenting policy, please let us know through the Contact Page.
  • NEW: Due to the recent increased amount of spam all links will require moderator approval. 
Hide Comment Policy
Comments