GIGABYTE AORUS RTX 5090 XTREME WATERFORCE WB vs NVIDIA RTX 4000 SFF Ada Generation

Comparison of GIGABYTE AORUS RTX 5090 XTREME WATERFORCE WB with 32 GB GDDR7 and 21,760 cores vs NVIDIA RTX 4000 SFF Ada Generation with 20 GB GDDR6 and 6,144 cores.

Loading...

Performance Rating

GIGABYTE AORUS RTX 5090 XTREME WATERFORCE WB outperforms NVIDIA RTX 4000 SFF Ada Generation by 369.29% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

GIGABYTE AORUS RTX 5090 XTREME WATERFORCE WB

46.5

GIGABYTE AORUS RTX 5090 XTREME WATERFORCE WB

46.5
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA RTX 4000 SFF Ada Generation

9.9

NVIDIA RTX 4000 SFF Ada Generation

9.9

Contents:

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

Memory

Memory Size

🔥 +60% 32 ГБ
20 ГБ

Memory Type

GDDR7 GDDR6

Memory Bandwidth

🔥 1.79 TB/s
280.0 GB/s

Memory Bus Width

512 бит 160 бит

ML Performance

FP16 (Half Precision)

🔥 +503% 115.5 TFLOPS
19.17 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +503% 115.5 TFLOPS
19.17 TFLOPS

FP64 (Double Precision)

🔥 +503% 1.805 TFLOPS
0.2995 TFLOPS

CUDA Cores

🔥 +254% 21,760
6,144

RT Cores

🔥 +254% 170
48

Architecture & Compatibility

GPU Architecture

Blackwell 2.0 Ada Lovelace

SM (Streaming Multiprocessor)

🔥 +254% 170
48

PCIe Version

PCIe 5.0 x16 PCIe 4.0 x16

ML Software Support

CUDA Version

🔥 12.0
8.9

Clocks & Performance

Base Clock

🔥 +180% 2,017
720

Boost Clock

🔥 +70% 2,655
1,560

Memory Clock

1,750 1,750

Power Consumption

TDP/TGP

575 W
🔥 -88% 70 W

Recommended PSU

950 W
🔥 -74% 250 W

Power Connector

1x 16-pin None

Rendering

Texture Units (TMU)

🔥 +254% 680
192

ROP

🔥 +254% 170
48

L2 Cache

🔥 +100% 96 MB
48 MB

Additional

Slots

Dual-slot Dual-slot

Release Date

Jan. 30, 2025 March 21, 2023

Display Outputs

1x HDMI 2.1b
3x DisplayPort 2.1b
4x mini-DisplayPort 1.4a

Renting is cheaper than buying