NVIDIA A100 PCIe 80 GB vs NVIDIA Tesla M40

Comparison of NVIDIA A100 PCIe 80 GB with 80 GB HBM2e and 6,912 cores vs NVIDIA Tesla M40 with 12 GB GDDR5 and 3,072 cores.

Loading...

Performance Rating

NVIDIA A100 PCIe 80 GB outperforms NVIDIA Tesla M40 by 348.86% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

NVIDIA A100 PCIe 80 GB

25.6

NVIDIA A100 PCIe 80 GB

25.6
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA Tesla M40

5.7

NVIDIA Tesla M40

5.7

Contents:

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

Memory

Memory Size

🔥 +567% 80 ГБ
12 ГБ

Memory Type

HBM2e GDDR5

Memory Bandwidth

🔥 1.94 TB/s
288.4 GB/s

Memory Bus Width

5,120 бит 384 бит

ML Performance

FP16 (Half Precision)

🔥 77.97 TFLOPS
No

BF16 (Brain Float)

🔥 311.84 TFLOPS
No

TF32 (TensorFloat)

🔥 155.92
No

Compute Power

FP32 (Single Precision)

🔥 +185% 19.49 TFLOPS
6.832 TFLOPS

FP64 (Double Precision)

🔥 +4,465% 9.746 TFLOPS
0.2135 TFLOPS

CUDA Cores

🔥 +125% 6,912
3,072

RT Cores

No No

Architecture & Compatibility

GPU Architecture

Ampere Maxwell 2.0

SM (Streaming Multiprocessor)

🔥 108
No

PCIe Version

PCIe 4.0 x16 PCIe 3.0 x16

ML Software Support

CUDA Version

🔥 8.0
5.2

Clocks & Performance

Base Clock

🔥 +12% 1,065
948

Boost Clock

🔥 +27% 1,410
1,112

Memory Clock

🔥 +1% 1,512
1,502

Power Consumption

TDP/TGP

300 W
🔥 -17% 250 W

Recommended PSU

700 W
🔥 -14% 600 W

Power Connector

8-pin EPS 8-pin EPS

Rendering

Texture Units (TMU)

🔥 +125% 432
192

ROP

No No

L2 Cache

🔥 +2,567% 80 MB
3 MB

Benchmarks

llama.cpp, gpt-oss 20B Q4_K - Medium

47.0 tokens/s

llama.cpp, llama 7B Q4_0

36.7 tokens/s

llama.cpp, llama-2-7b-Q4_0

41.7 tokens/s

llama.cpp, qwen3 32B Q4_K - Medium

7.19 tokens/s

llama.cpp, qwen3moe 30B.A3B Q4_K - Medium

35.1 tokens/s

Additional

Slots

Dual-slot Dual-slot

Release Date

June 28, 2021 Nov. 10, 2015

Display Outputs

No outputs
No outputs

Renting is cheaper than buying