AMD Radeon Instinct MI250 vs NVIDIA CMP 50HX

Comparison of AMD Radeon Instinct MI250 with 128 GB HBM2e and 13,312 cores vs NVIDIA CMP 50HX with 10 GB GDDR6 and 3,584 cores.

Loading...

Performance Rating

AMD Radeon Instinct MI250 outperforms NVIDIA CMP 50HX by 525.95% 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 CMP 50HX

8.1

NVIDIA CMP 50HX

8.1

Contents:

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

Memory

Memory Size

🔥 +1,180% 128 ГБ
10 ГБ

Memory Type

HBM2e GDDR6

Memory Bandwidth

🔥 3.28 TB/s
560.0 GB/s

Memory Bus Width

8,192 бит 320 бит

ML Performance

FP16 (Half Precision)

🔥 +1,535% 362.1 TFLOPS
22.15 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +309% 45.26 TFLOPS
11.07 TFLOPS

FP64 (Double Precision)

🔥 +12,977% 45.26 TFLOPS
0.3461 TFLOPS

CUDA Cores

🔥 +271% 13,312
3,584

RT Cores

No
🔥 56

Architecture & Compatibility

GPU Architecture

CDNA 2.0 Turing

SM (Streaming Multiprocessor)

No
🔥 56

PCIe Version

PCIe 4.0 x16 PCIe 1.0 x4

ML Software Support

CUDA Version

No 7.5

Clocks & Performance

Base Clock

1,000
🔥 +35% 1,350

Boost Clock

🔥 +10% 1,700
1,545

Memory Clock

1,600
🔥 +9% 1,750

Power Consumption

TDP/TGP

500 W
🔥 -50% 250 W

Recommended PSU

900 W
🔥 -33% 600 W

Power Connector

2x 8-pin 2x 8-pin

Rendering

Texture Units (TMU)

🔥 +333% 832
192

ROP

No
🔥 56

L2 Cache

16 MB
🔥 5 MB

Benchmarks

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

63.9 tokens/s

Additional

Slots

Dual-slot Dual-slot

Release Date

Nov. 8, 2021 June 24, 2021

Display Outputs

No outputs
No outputs

Renting is cheaper than buying