AMD Radeon Instinct MI355X vs Manli White Gallardo RTX 5090 OC

Comparison AMD Radeon Instinct MI355X with 288 GB HBM3e and 16,384 cores vs Manli White Gallardo RTX 5090 OC with 32 GB GDDR7 and 21,760 cores.

Loading...

Performance Rating

AMD Radeon Instinct MI355X outperforms Manli White Gallardo RTX 5090 OC by 108.84% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

AMD Radeon Instinct MI355X

93.5

AMD Radeon Instinct MI355X

93.5
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

Manli White Gallardo RTX 5090 OC

44.8

Manli White Gallardo RTX 5090 OC

44.8

Expert Comparison

AMD Radeon Instinct MI355X имеет значительно большую память (288 ГБ HBM3e против 32 ГБ GDDR7) и пропускную способность (8.19 TB/s против 1.79 TB/s), что делает его более эффективным для вычислительных задач, таких как машинное обучение и симуляции. Однако Manli White Gallardo RTX 5090 OC предлагает больше ядер (21760 против 16384) и FP32 вычислительной мощности (108.0 TFLOPS против 78.64 TFLOPS), что делает его предпочтительным для графики и игр.

Contents:

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

Memory

Memory Size

🔥 +800% 288 ГБ
32 GB

Memory Type

HBM3e GDDR7

Memory Bandwidth

🔥 +358% 8.19 TB/s
1.79 TB/s

Memory Bus Width

8,192 бит 512 бит

ML Performance

FP16 (Half Precision)

🔥 +482% 629.1 TFLOPS
108.0 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

78.64 TFLOPS
🔥 +37% 108.0 TFLOPS

FP64 (Double Precision)

🔥 +4,559% 78.64 TFLOPS
1.688 TFLOPS

CUDA Cores

16,384
🔥 +33% 21,760

RT Cores

No
🔥 170

Architecture & Compatibility

GPU Architecture

CDNA 4.0 Blackwell 2.0

SM (Streaming Multiprocessor)

No
🔥 170

PCIe Version

PCIe 5.0 x16 PCIe 5.0 x16

ML Software Support

CUDA Version

No 12.0

CUDA Toolkit (first supported)

v12

CUDA Toolkit status

Supported Supported

Clocks & Performance

Base Clock

1,000
🔥 +102% 2,017

Boost Clock

2,400
🔥 +3% 2,482

Memory Clock

🔥 +14% 2,000
1,750

Power Consumption

Recommended PSU

1800 W
🔥 -47% 950 W

Power Connector

None 1x 16-pin

TDP/TGP

1400 W
🔥 -59% 575 W

Rendering

Texture Units (TMU)

🔥 +51% 1,024
680

ROP

No
🔥 170

L2 Cache

32 MB
🔥 +200% 96 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 Triple-slot

Release Date

March 12, 2025 Jan. 30, 2025

Display Outputs

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

Renting is cheaper than buying