NVIDIA H100 PCIe 80 GB vs NVIDIA RTX PRO 4000 Blackwell SFF

Comparison of NVIDIA H100 PCIe 80 GB with 80 GB HBM2e and 14,592 cores vs NVIDIA RTX PRO 4000 Blackwell SFF with 24 GB GDDR7 and 8,960 cores.

Loading...

Performance Rating

NVIDIA H100 PCIe 80 GB outperforms NVIDIA RTX PRO 4000 Blackwell SFF by 231.95% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

NVIDIA H100 PCIe 80 GB

44.1

NVIDIA H100 PCIe 80 GB

44.1
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA RTX PRO 4000 Blackwell SFF

13.3

NVIDIA RTX PRO 4000 Blackwell SFF

13.3

Contents:

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

Memory

Memory Size

🔥 +233% 80 ГБ
24 GB

Memory Type

HBM2e GDDR7

Memory Bandwidth

🔥 2.04 TB/s
432.0 GB/s

Memory Bus Width

5,120 бит 192 бит

ML Performance

FP16 (Half Precision)

🔥 +699% 204.9 TFLOPS
25.66 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +100% 51.22 TFLOPS
25.66 TFLOPS

FP64 (Double Precision)

🔥 +6,287% 25.61 TFLOPS
0.401 TFLOPS

CUDA Cores

🔥 +63% 14,592
8,960

RT Cores

No
🔥 70

Architecture & Compatibility

GPU Architecture

Hopper Blackwell 2.0

SM (Streaming Multiprocessor)

🔥 +63% 114
70

PCIe Version

PCIe 5.0 x16 PCIe 5.0 x8

ML Software Support

CUDA Version

9.0
🔥 12.0

Clocks & Performance

Base Clock

🔥 +90% 1,095
577

Boost Clock

🔥 +23% 1,755
1,432

Memory Clock

🔥 +42% 1,593
1,125

Power Consumption

Recommended PSU

750 W
🔥 -67% 250 W

Power Connector

1x 16-pin None

TDP/TGP

350 W
🔥 -80% 70 W

Rendering

Texture Units (TMU)

🔥 +63% 456
280

ROP

No
🔥 70

L2 Cache

🔥 +4% 50 MB
48 MB

Benchmarks

MLPerf, llama2-70b-99.9 (fp8)

1 222 tokens/s

MLPerf, mixtral-8x7b (fp8)

3 466 tokens/s

Additional

Slots

Dual-slot Dual-slot

Release Date

Oct. 13, 2022 Aug. 11, 2025

Display Outputs

No outputs
4x mini-DisplayPort 2.1b

Renting is cheaper than buying