NVIDIA A40 PCIe vs NVIDIA B300

Comparison of NVIDIA A40 PCIe with 48 GB GDDR6 and 10,752 cores vs NVIDIA B300 with 144 GB HBM3e and 18,944 cores.

Loading...

Performance Rating

NVIDIA B300 outperforms NVIDIA A40 PCIe by 236.59% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

NVIDIA A40 PCIe

21.4

NVIDIA A40 PCIe

21.4
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA B300

71.9

NVIDIA B300

71.9

Contents:

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

Memory

Memory Size

48 GB
🔥 +200% 144 ГБ ×2 (288 ГБ)

Memory Type

GDDR6 HBM3e

Memory Bandwidth

695.8 GB/s
🔥 4.10 TB/s ×2 (8.2 TB/s)

Memory Bus Width

384 бит 4,096 бит ×2 (8192 бит)

ML Performance

FP16 (Half Precision)

🔥 +2,938% 37.42 TFLOPS
1.2318 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

37.42 TFLOPS
🔥 +106% 76.99 TFLOPS

FP64 (Double Precision)

0.5846 TFLOPS
🔥 +106% 1.2029 TFLOPS

CUDA Cores

10,752
🔥 +76% 18,944 ×2 (37888)

RT Cores

🔥 84
No

Architecture & Compatibility

GPU Architecture

Ampere Blackwell Ultra

SM (Streaming Multiprocessor)

84
🔥 +76% 148

PCIe Version

PCIe 4.0 x16 PCIe 5.0 x16

ML Software Support

CUDA Version

8.6
🔥 10.3

Clocks & Performance

Base Clock

1,305
🔥 +28% 1,665

Boost Clock

1,740
🔥 +17% 2,032

Memory Clock

1,812
🔥 +10% 2,000

Power Consumption

Recommended PSU

🔥 -61% 700 W
1800 W

Power Connector

8-pin EPS No

TDP/TGP

🔥 -79% 300 W
1400 W

Rendering

Texture Units (TMU)

336
🔥 +76% 592 ×2 (1184)

ROP

🔥 84
No

L2 Cache

🔥 6 MB
50 MB

Additional

Slots

Dual-slot
🔥 SXM Module

Release Date

Oct. 5, 2020 Sept. 11, 2025

Display Outputs

3x DisplayPort 1.4a
No outputs

Renting is cheaper than buying