AMD Radeon Instinct MI250 vs NVIDIA GeForce RTX 4090 Ti

Comparison of AMD Radeon Instinct MI250 with 128 GB HBM2e and 13,312 cores vs NVIDIA GeForce RTX 4090 Ti with 24 GB GDDR6X and 18,176 cores.

Loading...

Performance Rating

AMD Radeon Instinct MI250 outperforms NVIDIA GeForce RTX 4090 Ti by 44.66% 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 GeForce RTX 4090 Ti

35.2

NVIDIA GeForce RTX 4090 Ti

35.2

Contents:

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

Memory

Memory Size

🔥 +433% 128 ГБ
24 ГБ

Memory Type

HBM2e GDDR6X

Memory Bandwidth

🔥 +185% 3.28 TB/s
1.15 TB/s

Memory Bus Width

8,192 бит 384 бит

ML Performance

FP16 (Half Precision)

🔥 +288% 362.1 TFLOPS
93.24 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 45.26 TFLOPS
93.24 TFLOPS

FP64 (Double Precision)

🔥 45.26 TFLOPS
No

CUDA Cores

🔥 13,312
18,176

RT Cores

No 142

Architecture & Compatibility

GPU Architecture

CDNA 2.0 Ada Lovelace

SM (Streaming Multiprocessor)

No 142

PCIe Version

PCIe 4.0 x16 PCIe 4.0 x16

ML Software Support

CUDA Version

No 8.9

Clocks & Performance

Base Clock

🔥 1,000
2,325

Boost Clock

🔥 1,700
2,565

Memory Clock

🔥 +7% 1,600
1,500

Power Consumption

TDP/TGP

🔥 -17% 500 W
600 W

Recommended PSU

🔥 -10% 900 W
1000 W

Power Connector

2x 8-pin
🔥 2x 16-pin

Rendering

Texture Units (TMU)

🔥 +46% 832
568

ROP

No 142

L2 Cache

🔥 16 MB
96 MB

Benchmarks

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

63.9 tokens/s

Additional

Slots

Dual-slot Triple-slot

Release Date

Nov. 8, 2021 No

Display Outputs

No outputs
1x HDMI 2.1
3x DisplayPort 1.4a

Renting is cheaper than buying