Advantech RTX A1000 Embedded vs AMD Radeon Instinct MI325X

Comparison of Advantech RTX A1000 Embedded with 4 GB GDDR6 and 2,048 cores vs AMD Radeon Instinct MI325X with 288 GB HBM3e and 19,456 cores.

Loading...

Performance Rating

H200 H200
MI325X MI325X
A100 A100

Advantech RTX A1000 Embedded

Advantech RTX A1000 Embedded

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

AMD Radeon Instinct MI325X

AMD Radeon Instinct MI325X

Contents:

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

Memory

Memory Size

4 GB 288 GB

Memory Type

GDDR6 HBM3e

Memory Bandwidth

224.0 GB/s 10.3 TB/s

Memory Bus Width

128 бит 8,192 бит

ML Performance

FP16 (Half Precision)

6.664 TFLOPS 653.7 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

6.664 TFLOPS 81.72 TFLOPS

FP64 (Double Precision)

0.1041 TFLOPS 81.72 TFLOPS

CUDA Cores

2,048 19,456

RT Cores

16 No

Architecture & Compatibility

GPU Architecture

Ampere CDNA 3.0

SM (Streaming Multiprocessor)

16 No

PCIe Version

PCIe 4.0 x8 PCIe 5.0 x16

ML Software Support

CUDA Version

8.6 No

Clocks & Performance

Base Clock

1,192 1,000

Boost Clock

1,627 2,100

Memory Clock

1,750 2,525

Power Consumption

TDP/TGP

🔥 -94% 60 W
1000 W

Recommended PSU

No 1400 W

Power Connector

None None

Rendering

Texture Units (TMU)

64 1,216

ROP

16 No

L2 Cache

2 MB 16 MB

Benchmarks

MLPerf, llama2-70b-99.9 (Dummy)

3 596 tokens/s

MLPerf, llama2-70b-99.9 (fp8)

1 946 tokens/s

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

22.4 tokens/s

MLPerf, mixtral-8x7b (fp8)

6 975 tokens/s

Additional

Slots

MXM Module OAM Module

Release Date

March 30, 2022 Oct. 12, 2024

Display Outputs

Portable Device Dependent
No outputs

Renting is cheaper than buying