EVGA RTX 4090 FTW3 Prototype vs NVIDIA CMP 70HX

Comparison of EVGA RTX 4090 FTW3 Prototype with 24 GB GDDR6X and 16,384 cores vs NVIDIA CMP 70HX with 8 GB GDDR6X and 3,840 cores.

Loading...

Performance Rating

EVGA RTX 4090 FTW3 Prototype outperforms NVIDIA CMP 70HX by 317.99% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

EVGA RTX 4090 FTW3 Prototype

32.1

EVGA RTX 4090 FTW3 Prototype

32.1
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA CMP 70HX

7.7

NVIDIA CMP 70HX

7.7

Contents:

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

Memory

Memory Size

24 ГБ
🔥 8 ГБ

Memory Type

GDDR6X GDDR6X

Memory Bandwidth

🔥 1.01 TB/s
608.3 GB/s

Memory Bus Width

384 бит 256 бит

ML Performance

FP16 (Half Precision)

82.58 TFLOPS
🔥 10.71 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

82.58 TFLOPS
🔥 10.71 TFLOPS

FP64 (Double Precision)

No
🔥 0.1674 TFLOPS

CUDA Cores

16,384
🔥 3,840

RT Cores

128
🔥 30

Architecture & Compatibility

GPU Architecture

Ada Lovelace Ampere

SM (Streaming Multiprocessor)

128
🔥 30

PCIe Version

PCIe 4.0 x16 PCIe 1.0 x4

ML Software Support

CUDA Version

🔥 8.9
8.6

Clocks & Performance

Base Clock

2,235
🔥 1,365

Boost Clock

2,520
🔥 1,395

Memory Clock

1,313
🔥 1,188

Power Consumption

TDP/TGP

450 W unknown

Recommended PSU

850 W
🔥 -76% 200 W

Power Connector

1x 16-pin 1x 12-pin

Rendering

Texture Units (TMU)

512
🔥 120

ROP

128
🔥 30

L2 Cache

72 MB
🔥 4 MB

Additional

Slots

Quad-slot Dual-slot

Release Date

No July 16, 2021

Display Outputs

1x HDMI 2.1
3x DisplayPort 1.4a
No outputs

Renting is cheaper than buying