AMD Radeon Instinct MI250 vs NVIDIA Switch GPU 16nm

Comparison of AMD Radeon Instinct MI250 with 128 GB HBM2e and 13,312 cores vs NVIDIA Switch GPU 16nm with 4 GB DDR4 and 256 cores.

Loading...

Performance Rating

AMD Radeon Instinct MI250 outperforms NVIDIA Switch GPU 16nm by 119.64% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

AMD Radeon Instinct MI250

50.9

AMD Radeon Instinct MI250

50.9
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA Switch GPU 16nm

23.2

NVIDIA Switch GPU 16nm

23.2

Contents:

Memory ML Performance Compute Power Architecture & Compatibility ML Software Support Clocks & Performance Power Consumption Rendering Benchmarks Additional

Memory

Memory Size

🔥 +3,100% 128 ГБ
4 ГБ

Memory Type

HBM2e DDR4

Memory Bandwidth

🔥 3.28 TB/s
25.60 GB/s

Memory Bus Width

8,192 бит 64 бит

ML Performance

FP16 (Half Precision)

🔥 362.1 TFLOPS
786.4 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 45.26 TFLOPS
393.2 TFLOPS

FP64 (Double Precision)

🔥 +368,167% 45.26 TFLOPS
0.012289999999999999 TFLOPS

CUDA Cores

🔥 +5,100% 13,312
256

RT Cores

No No

Architecture & Compatibility

GPU Architecture

CDNA 2.0 Maxwell 2.0

SM (Streaming Multiprocessor)

No No

PCIe Version

PCIe 4.0 x16 No

ML Software Support

CUDA Version

No 5.3

Clocks & Performance

Base Clock

🔥 +160% 1,000
384

Boost Clock

🔥 +121% 1,700
768

Memory Clock

🔥 1,600
1,600

Power Consumption

TDP/TGP

500 W
🔥 -97% 15 W

Recommended PSU

900 W No

Power Connector

2x 8-pin No

Rendering

Texture Units (TMU)

🔥 +5,100% 832
16

ROP

No No

L2 Cache

🔥 16 MB
No

Benchmarks

llama.cpp, llama-2-7b-Q4_0

63.9 tokens/s

Additional

Slots

Dual-slot No

Release Date

Nov. 8, 2021 Aug. 16, 2019

Display Outputs

No outputs
1x USB Type-C

Renting is cheaper than buying