AMD Radeon Instinct MI250 vs NVIDIA A10M

Comparison of AMD Radeon Instinct MI250 with 128 GB HBM2e and 13,312 cores vs NVIDIA A10M with 20 GB GDDR6 and 7,168 cores.

Loading...

Performance Rating

H200 H200
MI325X MI325X
A100 A100

AMD Radeon Instinct MI250

50.9

AMD Radeon Instinct MI250

50.9
MI250 MI250
Instinct MI300X Instinct MI300X
RX 7900 XTX RX 7900 XTX

NVIDIA A10M

NVIDIA A10M

Contents:

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

Memory

Memory Size

🔥 +540% 128 ГБ
20 GB

Memory Type

HBM2e GDDR6

Memory Bandwidth

🔥 3.28 TB/s
500.2 GB/s

Memory Bus Width

8,192 бит 320 бит

ML Performance

FP16 (Half Precision)

🔥 +1,445% 362.1 TFLOPS
23.44 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +93% 45.26 TFLOPS
23.44 TFLOPS

FP64 (Double Precision)

🔥 +6,079% 45.26 TFLOPS
0.7325 TFLOPS

CUDA Cores

🔥 +86% 13,312
7,168

RT Cores

No 56

Architecture & Compatibility

GPU Architecture

CDNA 2.0 Ampere

SM (Streaming Multiprocessor)

No 56

PCIe Version

PCIe 4.0 x16 PCIe 4.0 x16

ML Software Support

CUDA Version

No 8.6

Clocks & Performance

Base Clock

🔥 +3% 1,000
975

Boost Clock

🔥 +4% 1,700
1,635

Memory Clock

🔥 +2% 1,600
1,563

Power Consumption

TDP/TGP

500 W
🔥 -70% 150 W

Recommended PSU

900 W
🔥 -50% 450 W

Power Connector

2x 8-pin 8-pin EPS

Rendering

Texture Units (TMU)

🔥 +271% 832
224

ROP

No 56

L2 Cache

🔥 +167% 16 MB
6 MB

Benchmarks

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

63.9 tokens/s

Additional

Slots

Dual-slot Single-slot

Release Date

Nov. 8, 2021 Feb. 11, 2022

Display Outputs

No outputs
No outputs

Renting is cheaper than buying