AMD Radeon Instinct MI355X vs NVIDIA A100 PCIe 40 GB

Comparison of AMD Radeon Instinct MI355X with 288 GB HBM3e and 16,384 cores vs NVIDIA A100 PCIe 40 GB with 40 GB HBM2e and 6,912 cores.

Loading...

Performance Rating

H200 H200
MI325X MI325X
A100 A100

AMD Radeon Instinct MI355X

AMD Radeon Instinct MI355X

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

NVIDIA A100 PCIe 40 GB

NVIDIA A100 PCIe 40 GB

Contents:

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

Memory

Memory Size

288 GB 40 GB

Memory Type

HBM3e HBM2e

Memory Bandwidth

8.19 TB/s 1.56 TB/s

Memory Bus Width

8,192 бит 5,120 бит

ML Performance

FP16 (Half Precision)

629.1 TFLOPS 77.97 TFLOPS

BF16 (Brain Float)

No 311.84 TFLOPS

TF32 (TensorFloat)

No 155.92

Compute Power

FP32 (Single Precision)

78.64 TFLOPS 19.49 TFLOPS

FP64 (Double Precision)

78.64 TFLOPS 9.746 TFLOPS

CUDA Cores

16,384 6,912

RT Cores

No No

Architecture & Compatibility

GPU Architecture

CDNA 4.0 Ampere

SM (Streaming Multiprocessor)

No 108

PCIe Version

PCIe 5.0 x16 PCIe 4.0 x16

ML Software Support

CUDA Version

No 8.0

Clocks & Performance

Base Clock

1,000 765

Boost Clock

2,400 1,410

Memory Clock

2,000 1,215

Power Consumption

TDP/TGP

1400 W
🔥 -82% 250 W

Recommended PSU

1800 W
🔥 -67% 600 W

Power Connector

None 8-pin EPS

Rendering

Texture Units (TMU)

1,024 432

ROP

No No

L2 Cache

32 MB 40 MB

Benchmarks

MLPerf, llama2-70b-99.9 (mxfp4)

10 129 tokens/s

MLPerf, llama3.1-405b (mxfp4)

252.4 tokens/s

Additional

Slots

OAM Module Dual-slot

Release Date

March 12, 2025 June 22, 2020

Display Outputs

No outputs
No outputs

Renting is cheaper than buying