NVIDIA B300 vs NVIDIA ION LE

Comparison NVIDIA B300 with 144 GB HBM3e and 18,944 cores vs NVIDIA ION LE and 16 cores.

Loading...

Performance Rating

A100 A100
H200 H200
MI325X MI325X

NVIDIA B300

71.9

NVIDIA B300

71.9
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA ION LE

NVIDIA ION LE

Expert Comparison

NVIDIA B300 значительно превосходит NVIDIA ION LE в практически всех аспектах: от количества ядер и пропускной способности до TDP и GPU Performance Index. B300 идеально подходит для высокопроизводительных задач, таких как машинное обучение и визуализация, в то время как ION LE более подходящ для легких задач или тестирования.

Contents:

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

Memory

Memory Size

🔥 144 ГБ ×2 (288 ГБ)
No

Memory Type

HBM3e System Shared

Memory Bandwidth

🔥 4.10 TB/s ×2 (8.2 TB/s)
System Dependent

Memory Bus Width

4,096 бит ×2 (8192 бит) No

ML Performance

FP16 (Half Precision)

🔥 1.2318 TFLOPS
No

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +218,622% 76.99 TFLOPS
0.0352 TFLOPS

FP64 (Double Precision)

🔥 1.2029 TFLOPS
No

CUDA Cores

🔥 +118,300% 18,944 ×2 (37888)
16

RT Cores

No No

Architecture & Compatibility

GPU Architecture

Blackwell Ultra Tesla

SM (Streaming Multiprocessor)

🔥 +14,700% 148
1

PCIe Version

PCIe 5.0 x16 PCI

ML Software Support

CUDA Version

10.3 No

CUDA Toolkit (first supported)

v12 v1

CUDA Toolkit (deprecated)

v7

CUDA Toolkit (support removed)

v8

CUDA Toolkit status

Supported Removed

Clocks & Performance

Base Clock

🔥 1,665
No

Boost Clock

🔥 2,032
No

Memory Clock

🔥 2,000
No

Power Consumption

Recommended PSU

1800 W No

Power Connector

No No

TDP/TGP

1400 W
🔥 -99% 20 W

Rendering

Texture Units (TMU)

🔥 +7,300% 592 ×2 (1184)
8

ROP

No No

L2 Cache

🔥 50 MB
No

Additional

Slots

🔥 SXM Module
IGP

Release Date

Sept. 11, 2025 June 3, 2008

Display Outputs

No outputs
Portable Device Dependent

Renting is cheaper than buying