AMD Radeon Instinct MI250 vs NVIDIA GeForce RTX 5070 Ti Mobile

Comparison of AMD Radeon Instinct MI250 with 128 GB HBM2e and 13,312 cores vs NVIDIA GeForce RTX 5070 Ti Mobile with 12 GB GDDR7 and 6,400 cores.

Loading...

Performance Rating

H200 H200
MI325X MI325X
A100 A100

AMD Radeon Instinct MI250

AMD Radeon Instinct MI250

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

NVIDIA GeForce RTX 5070 Ti Mobile

14.5

NVIDIA GeForce RTX 5070 Ti Mobile

14.5

Contents:

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

Memory

Memory Size

128 GB
🔥 12 ГБ

Memory Type

HBM2e GDDR7

Memory Bandwidth

3.28 TB/s 608.6 GB/s

Memory Bus Width

8,192 бит 192 бит

ML Performance

FP16 (Half Precision)

362.1 TFLOPS
🔥 32.26 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

45.26 TFLOPS
🔥 32.26 TFLOPS

FP64 (Double Precision)

45.26 TFLOPS
🔥 +1,113,466% 504,000.0 TFLOPS

CUDA Cores

13,312
🔥 6,400

RT Cores

No
🔥 50

Architecture & Compatibility

GPU Architecture

CDNA 2.0 Blackwell 2.0

SM (Streaming Multiprocessor)

No
🔥 50

PCIe Version

PCIe 4.0 x16 PCIe 5.0 x16

ML Software Support

CUDA Version

No 10.1

Clocks & Performance

Base Clock

1,000
🔥 +124% 2,235

Boost Clock

1,700
🔥 +48% 2,520

Memory Clock

1,600
🔥 +25% 2,000

Power Consumption

TDP/TGP

500 W
🔥 -72% 140 W

Recommended PSU

900 W No

Power Connector

2x 8-pin None

Rendering

Texture Units (TMU)

832
🔥 200

ROP

No
🔥 50

L2 Cache

16 MB
🔥 +150% 40 MB

Benchmarks

Hashcat: MD5

90.2 GH/s

Hashcat: SHA-256

11.6 GH/s

Hashcat: SHA-512

4.11 GH/s

Blender: Blend File: Barbershop - Compute: NVIDIA CUDA

78.09 Seconds

Blender: Blend File: Barbershop - Compute: NVIDIA OptiX

46.33 Seconds

Blender: Blend File: Fishy Cat - Compute: NVIDIA CUDA

20.60 Seconds

Blender: Blend File: Fishy Cat - Compute: NVIDIA OptiX

8.45 Seconds

Blender: Blend File: Junkshop - Compute: NVIDIA CUDA

15.22 Seconds

Blender: Blend File: Junkshop - Compute: NVIDIA OptiX

8.28 Seconds

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

63.9 tokens/s

Additional

Slots

Dual-slot IGP

Release Date

Nov. 8, 2021 March 13, 2025

Display Outputs

No outputs
1x HDMI 2.1b
3x DisplayPort 2.1a

Renting is cheaper than buying