GPU Wiki

Compare GPUs

Pick up to four GPUs and compare their specs row by row. Specs that mean different things across generations are kept apart.

GPU 1

GPU 2

Add a third or fourth GPU

GPU 3 (optional)

GPU 4 (optional)

Spec	A100 PCIe 40 GB
Overview
Product name	NVIDIA A100 PCIe 40 GB
Architecture	Ampere
Generation / family	NVIDIA A-Series
GPU codename	GA100
Launch date	2020-06-22
Market segment	Data Center in NVIDIA A-Series
Board type / entry type	Data Center GPU
GPU
Process node	TSMC 7 nm
Die size	826 mm²
Transistor count	54.2 billion
GPU chip	GA100
SM / multiprocessor count	108
CUDA Cores	6912
Tensor Cores	432
ROPs	160
TMUs	432
Clocks
Base / Core clock	765 MHz
Boost clock	1,410 MHz
Memory clock	1,215 MHz
Memory
Memory size	40 GB
Memory type	HBM2
Memory bus width	5,120-bit
Memory bandwidth	1,560 GB/s
Bus / Interface
Bus interface	PCIe 4.0 x16
AGP / PCI / PCIe distinction	PCIe
Power
TDP / TBP / board power	250 W
Recommended PSU	600 W
Power connectors	1x 8-pin CPU (EPS-12V)
Slot width	Dual-slot, 267 mm length
Features
CUDA support	8.0
Performance / Benchmarks
FP32 throughput	19.49 TFLOPS
FP16 throughput	77.97 TFLOPS

Overview

Product name

A100 PCIe 40 GB NVIDIA A100 PCIe 40 GB

Architecture

A100 PCIe 40 GB Ampere

Generation / family

A100 PCIe 40 GB NVIDIA A-Series

GPU codename

A100 PCIe 40 GB GA100

Launch date

A100 PCIe 40 GB 2020-06-22

Market segment

A100 PCIe 40 GB Data Center in NVIDIA A-Series

Board type / entry type

A100 PCIe 40 GB Data Center GPU

GPU

Process node

A100 PCIe 40 GB TSMC 7 nm

Die size

A100 PCIe 40 GB 826 mm²

Transistor count

A100 PCIe 40 GB 54.2 billion

GPU chip

A100 PCIe 40 GB GA100

SM / multiprocessor count

A100 PCIe 40 GB 108

CUDA Cores

A100 PCIe 40 GB 6912

Tensor Cores

A100 PCIe 40 GB 432

ROPs

A100 PCIe 40 GB 160

TMUs

A100 PCIe 40 GB 432

Clocks

Base / Core clock

A100 PCIe 40 GB 765 MHz

Boost clock

A100 PCIe 40 GB 1,410 MHz

Memory clock

A100 PCIe 40 GB 1,215 MHz

Memory

Memory size

A100 PCIe 40 GB 40 GB

Memory type

A100 PCIe 40 GB HBM2

Memory bus width

A100 PCIe 40 GB 5,120-bit

Memory bandwidth

A100 PCIe 40 GB 1,560 GB/s

Bus / Interface

Bus interface

A100 PCIe 40 GB PCIe 4.0 x16

AGP / PCI / PCIe distinction

A100 PCIe 40 GB PCIe

Power

TDP / TBP / board power

A100 PCIe 40 GB 250 W

Recommended PSU

A100 PCIe 40 GB 600 W

Power connectors

A100 PCIe 40 GB 1x 8-pin CPU (EPS-12V)

Slot width

A100 PCIe 40 GB Dual-slot, 267 mm length

Features

CUDA support

A100 PCIe 40 GB 8.0

Performance / Benchmarks

FP32 throughput

A100 PCIe 40 GB 19.49 TFLOPS

FP16 throughput

A100 PCIe 40 GB 77.97 TFLOPS

Muted cells indicate unavailable, not applicable, or not directly comparable specifications.

DLAA counts as native resolution; DLSS, FSR and XeSS upscaling never do. A ray tracing setting without a readable on or off value is read as on.