Home NVIDIA A40 PCIe

NVIDIA A40 PCIe

This is a GPU manufactured with Samsung 8nm process, based on Nvidia Ampere architecture and released on Oct 2020. It features 28.3 billion transistors, 10752 CUDA cores and 48GB GDDR6 memory, with 6MB L2 cache, theoretical performance of 37.42TFLOPS, with total power consumption of 300W.

Graphics Card

[Report Issues]
Release Date
Oct 2020
Generation
Tesla
Type
Desktop
Bus Interface
PCIe 4.0 x16

Clock Speeds

[Report Issues]
Base Clock
1305MHz
Boost Clock
1740MHz
Memory Clock
1812MHz

Memory

[Report Issues]
Memory Size
48GB
Memory Type
GDDR6
Memory Bus
384bit
Bandwidth
695.8 GB/s

Render Config

[Report Issues]
SM Count
84
Shading Units
10752
TMUs
336
ROPs
112
Tensor Cores
336
RT Cores
84
L1 Cache
128 KB (per SM)
L2 Cache
6MB

Theoretical Performance

[Report Issues]
Pixel Rate
194.9 GPixel/s
Texture Rate
584.6 GTexel/s
FP16 (half)
37.42 TFLOPS
FP32 (float)
37.42 TFLOPS
FP64 (double)
584.6 GFLOPS

Graphics Processor

[Report Issues]
GPU Name
GA102
Architecture
Ampere
Foundry
Samsung
Process Size
8 nm
Transistors
28,300 million
Die Size
628mm²

Board Design

[Report Issues]
TDP
300W
Suggested PSU
700W
Outputs
3x DisplayPort 1.4a
Power Connectors
8-pin EPS

Graphics Features

[Report Issues]
DirectX
12 Ultimate (12_2)
OpenGL
4.6
OpenCL
3.0
Vulkan
1.3
CUDA
8.6
Shader Model
6.6

Rankings

[Report Issues]
FP32 (float)
39.63 TFLOPS
NVIDIA RTX A6000 48 GB GDDR6
38.71 TFLOPS
38.52 TFLOPS
NVIDIA A40 PCIe 48 GB GDDR6
37.42 TFLOPS
37.32 TFLOPS
AMD Radeon RX 7800 16 GB GDDR6
37.32 TFLOPS
35.58 TFLOPS
Blender
NVIDIA H100 PCIe 80 GB HBM2e
4845
4310
NVIDIA A40 PCIe 48 GB GDDR6
4170
4164

Related Comparisons

© 2024 - TopCPU.net   Contact Us Privacy Policy