NVIDIA A40 PCIe vs NVIDIA RTX PRO 6000 Blackwell Server

Comparison of NVIDIA A40 PCIe with 48 GB GDDR6 and 10,752 cores vs NVIDIA RTX PRO 6000 Blackwell Server with 96 GB GDDR7 and 24,064 cores.

Loading...

Performance Rating

NVIDIA RTX PRO 6000 Blackwell Server outperforms NVIDIA A40 PCIe by 168.93% 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 RTX PRO 6000 Blackwell Server

57.5

NVIDIA RTX PRO 6000 Blackwell Server

57.5

Expert Comparison

NVIDIA A40 PCIe имеет меньшее количество ядер и памяти по сравнению с NVIDIA RTX PRO 6000 Blackwell Server, но потребляет меньше энергии и обеспечивает более высокую пропускную способность (695.8 GB/s против 1.79 TB/s). RTX PRO 6000 Blackwell Server же обладает значительно большими вычислительными мощностями (126.0 TFLOPS против 37.42 TFLOPS) и большей памятью (96 ГБ против 48 ГБ), что делает его предпочтительным для интенсивных вычислений и больших объемов данных.

Contents:

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

Memory

Memory Size

48 ГБ
🔥 +100% 96 ГБ

Memory Type

GDDR6 GDDR7

Memory Bandwidth

695.8 GB/s
🔥 1.79 TB/s

Memory Bus Width

384 бит 512 бит

ML Performance

FP16 (Half Precision)

37.42 TFLOPS
🔥 +237% 126.0 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

37.42 TFLOPS
🔥 +237% 126.0 TFLOPS

FP64 (Double Precision)

0.5846 TFLOPS
🔥 +237% 1.968 TFLOPS

CUDA Cores

10,752
🔥 +124% 24,064

RT Cores

84
🔥 +124% 188

Architecture & Compatibility

GPU Architecture

Ampere Blackwell 2.0

SM (Streaming Multiprocessor)

84
🔥 +124% 188

PCIe Version

PCIe 4.0 x16 PCIe 5.0 x16

ML Software Support

CUDA Version

8.6
🔥 12.0

Clocks & Performance

Base Clock

1,305
🔥 +22% 1,590

Boost Clock

1,740
🔥 +50% 2,617

Memory Clock

🔥 +4% 1,812
1,750

Power Consumption

TDP/TGP

🔥 -50% 300 W
600 W

Recommended PSU

🔥 -30% 700 W
1000 W

Power Connector

8-pin EPS 1x 16-pin

Rendering

Texture Units (TMU)

336
🔥 +124% 752

ROP

84
🔥 +124% 188

L2 Cache

🔥 6 MB
128 MB

Benchmarks

MLPerf, llama2-70b-99.9 (fp4)

3 250 tokens/s

MLPerf, llama3.1-8b (fp4)

5 758 tokens/s

Geekbench AI, FP16

53 322 points

Geekbench AI, INT8

28 264 points

Geekbench AI, FP32

37 299 points

MLPerf, mixtral-8x7b (fp8)

3 767 tokens/s

Additional

Slots

Dual-slot Dual-slot

Release Date

Oct. 5, 2020 March 18, 2025

Display Outputs

3x DisplayPort 1.4a
4x DisplayPort 2.1b

Renting is cheaper than buying