AMD Radeon Instinct MI250 vs AMD Radeon R9 M395X

Comparison AMD Radeon Instinct MI250 with 128 GB HBM2e and 13,312 cores vs AMD Radeon R9 M395X with 8 GB GDDR5 and 2,048 cores.

Loading...

Performance Rating

AMD Radeon Instinct MI250 outperforms AMD Radeon R9 M395X by 1,654.83% 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

AMD Radeon R9 M395X

2.9

AMD Radeon R9 M395X

2.9

Expert Comparison

AMD Radeon Instinct MI250 значительно выигрывает у Radeon R9 M395X по всем параметрам: имеет больше ядер, больший объем памяти, большую пропускную способность и гораздо более высокую производительность в FP32. MI250 предназначена для вычислительных задач и больших серверов, тогда как R9 M395X подходит для игр и轻量级任务。

Contents:

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

Memory

Memory Size

🔥 +1,500% 128 ГБ
8 GB

Memory Type

HBM2e GDDR5

Memory Bandwidth

🔥 3.28 TB/s
160.0 GB/s

Memory Bus Width

8,192 бит 256 бит

ML Performance

FP16 (Half Precision)

🔥 +12,129% 362.1 TFLOPS
2.961 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +1,429% 45.26 TFLOPS
2.961 TFLOPS

FP64 (Double Precision)

🔥 +24,352% 45.26 TFLOPS
0.1851 TFLOPS

CUDA Cores

🔥 +550% 13,312
2,048

RT Cores

No No

Architecture & Compatibility

GPU Architecture

CDNA 2.0 GCN 3.0

SM (Streaming Multiprocessor)

No No

PCIe Version

PCIe 4.0 x16 PCIe 3.0 x16

ML Software Support

CUDA Version

No No

CUDA Toolkit status

Supported Supported

Clocks & Performance

Base Clock

🔥 1,000
No

Boost Clock

🔥 1,700
No

Memory Clock

🔥 +28% 1,600
1,250

Power Consumption

Recommended PSU

900 W No

Power Connector

2x 8-pin None

TDP/TGP

500 W
🔥 -85% 75 W

Rendering

Texture Units (TMU)

🔥 +550% 832
128

ROP

No No

L2 Cache

🔥 16 MB
512 KB

Benchmarks

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

63.9 tokens/s

Additional

Slots

Dual-slot MXM Module

Release Date

Nov. 8, 2021 May 5, 2015

Display Outputs

No outputs
Portable Device Dependent

Renting is cheaper than buying