rent-a-gpu.eu

NVIDIA · Ampere

Rent A100

NVIDIA A100 Tensor Core GPU, 80 GB.general mlLLM inferenceHPC
Full price from
—
in stock, per GPU-hour
Spot from
—
in stock, per GPU-hour
Offers in stock
0 of 23
full price or spot
Providers in stock
0 of 6
full price or spot

Offers (24)

HyperstackNo stock dataSelf-service

1× A10080 GB VRAM

Canada · 28 vCPU · 120 GB RAM · 100GB root + 750GB ephemeral

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

n3-A100x1

Full price
€1.20
per hour, whole instance, 1 GPU
Spot €0.96/h · no stock data
View offer
HyperstackNo stock dataSelf-serviceNVLink 3 possible, not confirmed

2× A10080 GB each, 160 GB total

Canada · 60 vCPU · 240 GB RAM · 100GB root + 1500GB ephemeral

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.

n3-A100x2

Full price
€2.41
per hour, whole instance, 2 GPUs
€1.20/h per GPU
Spot €1.93/h instance · €0.96/h per GPU · no stock data
View offer
HyperstackNo stock dataSelf-serviceNVLink 3 possible, not confirmed

8× A10080 GB each, 640 GB total

Canada · 252 vCPU · 960 GB RAM · 100GB root + 6500GB ephemeral

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.

n3-A100x8

Full price
€9.63
per hour, whole instance, 8 GPUs
€1.20/h per GPU
Spot €7.70/h instance · €0.96/h per GPU · no stock data
View offer
HyperstackNo stock dataSelf-serviceNVLink 3 possible, not confirmed

4× A10080 GB each, 320 GB total

Canada · 124 vCPU · 480 GB RAM · 100GB root + 3200GB ephemeral

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.

n3-A100x4

Full price
€4.82
per hour, whole instance, 4 GPUs
€1.20/h per GPU
Spot €3.85/h instance · €0.96/h per GPU · no stock data
View offer
HyperstackNo stock dataSelf-serviceNVLink 3 possible, not confirmed

8× A10080 GB each, 640 GB total

Canada · 252 vCPU · 1920 GB RAM · 100GB root + 6500GB ephemeral

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.

n3-A100x8-NVLink

Full price
€9.99
per hour, whole instance, 8 GPUs
€1.25/h per GPU
View offer
GcoreNo stock dataSelf-service, ready in 15–40 minNVLink 3 possible, not confirmed

8× A10080 GB each, 640 GB total

Luxembourg · 2048 GB RAM · 8x3.84 TB NVMe

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.

bm3-ai-1xlarge-noib-a100-80-8

Full price
€10.91
per hour, whole instance, 8 GPUs
€1.36/h per GPU
View offer
HyperstackNo stock dataSelf-serviceLinked · NVLink 3 · 600 GB/s

8× A10080 GB each, 640 GB total

United States · 192 vCPU · 960 GB RAM · 100GB root + 16000GB ephemeral

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

The hardware type or the provider confirms a fast GPU-to-GPU link. Software still has to split the work across GPUs. They never appear as one big memory.

n3-A100-SXM4x8

Full price
€11.41
per hour, whole instance, 8 GPUs
€1.43/h per GPU
Spot €9.13/h instance · €1.14/h per GPU · no stock data
View offer
STACKITNo stock dataSelf-service

1× A10080 GB VRAM

Germany · 14 vCPU · 108 GB RAM

Details

Stock and price checked 14 min ago.

You start it yourself once there is stock.

n1.14d.g1

Full price
€2.66
per hour, whole instance, 1 GPU
View offer
Provider not namedMight be available

A100

Non-EU · 1 offer

Full price
from €1.20
per GPU, per hour
See all 23 offers

Price and availability history

Price per GPU-hour

Excluding VAT.

Providers with stock

Each provider counts once. Full price and spot are separate.

Share of offers in stock

Full price. Offers we can't say are really sold there are left out until they've been in stock once.

Units free

Only providers that report a number. Each line is one provider and one server size, because their counts cover different things and can't be added up. Providers that only say yes or no are under Stock by provider.

Stock by provider

In stock, yes or no, for every provider selling this GPU. A filled band means in stock: green at full price, orange as spot.

NosanaSold out
Reports a number
CloudHaus.ioNo answer in the last check
GcoreNo stock data
No stock data shown
HyperstackNo stock data
No stock data shown
OVHcloudNo stock data
No stock data shown
STACKITNo stock data
No stock data shown
T Cloud PublicNo stock data
No stock data shown

Using several together

Depends on the version. The SXM version is fully linked with NVLink 3 at up to 600 GB/s. The PCIe version links in pairs at best, and only if the provider has fitted a bridge. Software still has to split the work, so the memory doesn't act as one big pool.

Specifications

Headline values are SXM4 80GB. 40GB variants use HBM2 at 1555 GB/s; PCIe 80GB bandwidth is 1935 GB/s. memoryDataRateGbps is derived for SXM4 80GB (2039 GB/s x 8 / 5120 bits). maxGpusInFabric is the 8-GPU NVSwitch domain NVIDIA documents for DGX/HGX A100. cooling is the PCIe card (passive); SXM module cooling is not stated.

Memory

VRAM
80 GB
Memory type
HBM2e
Memory bandwidth
How fast the GPU reaches its memory. Often limits LLM speed
2,039 GB/s
Bus width
5,120 bit
ECC memory
Detects and corrects memory errors
Yes

Compute

FP64
Speed at 64-bit precision, used for scientific simulation
9.7 TFLOPS
FP32
Speed at 32-bit precision
19.5 TFLOPS
FP16
Speed at 16-bit precision
624 TFLOPS
BF16
16-bit AI format. Standard for training and full-quality inference
624 TFLOPS
INT8
8-bit integer speed, used by quantized models
1,248 TOPS
FP8 support
No
FP4 support
No
Transformer Engine
Hardware that speeds up FP8 and FP4
No
Compute cores
6,912

Multi-GPU and connectivity

Interconnect
NVLink 3
Interconnect bandwidth
How fast linked GPUs exchange data
600 GB/s
Links
12
Per link
50 GB/s
Max GPUs in fabric
8
NVSwitch
Lets every GPU in a server talk to every other at full speed
Yes
PCIe generation
4
PCIe lanes
16
PCIe bandwidth
64 GB/s

Sharing and security

MIG partitioning
Splits one GPU into several isolated smaller GPUs
Yes
Max MIG instances
7
Confidential computing
No

Hardware

Architecture
Ampere
Launch year
2020
Process node
TSMC N7
Transistors
54.2 billion
Die size
826 mm²
Power (TDP)
Maximum power draw
400 W
Cooling
passive

Video

NVENC engines
0
NVDEC engines
5
AV1 encode
No
AV1 decode
No
Sources (46)