AMD Radeon Instinct MI250 vs NVIDIA Tesla V100 SXM2 16 GB

Comparison of AMD Radeon Instinct MI250 with 128 GB HBM2e and 13,312 cores vs NVIDIA Tesla V100 SXM2 16 GB with 16 GB HBM2 and 5,120 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 Tesla V100 SXM2 16 GB

16.0

NVIDIA Tesla V100 SXM2 16 GB

16.0

Contents:

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

Memory

Memory Size

128 GB 16 ГБ

Memory Type

HBM2e HBM2

Memory Bandwidth

3.28 TB/s
🔥 1.13 TB/s

Memory Bus Width

8,192 бит 4,096 бит

ML Performance

FP16 (Half Precision)

362.1 TFLOPS 32.71 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

45.26 TFLOPS 16.35 TFLOPS

FP64 (Double Precision)

45.26 TFLOPS 8.177 TFLOPS

CUDA Cores

13,312 5,120

RT Cores

No No

Architecture & Compatibility

GPU Architecture

CDNA 2.0 Volta

SM (Streaming Multiprocessor)

No 80

PCIe Version

PCIe 4.0 x16 PCIe 3.0 x16

ML Software Support

CUDA Version

No 7.0

Clocks & Performance

Base Clock

1,000 1,245

Boost Clock

1,700 1,597

Memory Clock

1,600 1,106

Power Consumption

TDP/TGP

500 W
🔥 -50% 250 W

Recommended PSU

900 W
🔥 -33% 600 W

Power Connector

2x 8-pin None

Rendering

Texture Units (TMU)

832 320

ROP

No No

L2 Cache

16 MB 6 MB

Benchmarks

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

63.9 tokens/s

Additional

Slots

Dual-slot
🔥 SXM Module

Release Date

Nov. 8, 2021 Nov. 26, 2019

Display Outputs

No outputs
No outputs

Renting is cheaper than buying