NVIDIA · Ada Lovelace

Rent RTX 4070

NVIDIA GeForce RTX 4070, 12 GB.inference smallRendering

Full price from

—

in stock, per GPU-hour

Spot from

—

in stock, per GPU-hour

Offers in stock

0 of 1

full price or spot

Providers in stock

0 of 1

full price or spot

Offers (1)

NosanaSold outSelf-service

1× RTX 407012 GB VRAM

Nosana network, hosts worldwide

Details

Stock and price checked 7 min ago.

You start it yourself once there is stock.

nvidia-4070

Full price

€0.09

per hour, whole instance, 1 GPU

View offer

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

Using several together

Can't be linked. This GPU has no NVLink, so several in one server only talk over PCIe. Each card keeps its own 12 GB. That suits running a separate job on each card. Splitting one model across them works, but slowly.

Specifications

Consumer card. Headline values are the original GDDR6X model; NVIDIA lists GDDR6 / GDDR6X for this card. The GDDR6 variant is 20 Gbps per TechPowerUp; its memoryBandwidthGBs is derived (192 bits x 20 Gbps / 8 = 480). Memory bandwidth and FLOPS figures come from NVIDIA's Blackwell whitepaper (Table 6), not a product datasheet. FP8 (FP16 accumulate), FP16 (FP16 accumulate), BF16 (FP32 accumulate) and INT8 are with sparsity. fp64Tflops stays null because NVIDIA states the 1/64th FP64 ratio only for AD102, and this card uses AD104. No NVLink/SLI.

Memory

VRAM: 12 GB
Memory type: GDDR6X
Memory bandwidth How fast the GPU reaches its memory. Often limits LLM speed: 504 GB/s
Bus width: 192 bit

Compute

FP32 Speed at 32-bit precision: 29.1 TFLOPS
FP16 Speed at 16-bit precision: 233.2 TFLOPS
BF16 16-bit AI format. Standard for training and full-quality inference: 116.6 TFLOPS
FP8 8-bit AI format. About half the memory of BF16, slightly lower quality: 466.4 TFLOPS
INT8 8-bit integer speed, used by quantized models: 466.4 TOPS
FP8 support: Yes
FP4 support: No
Transformer Engine Hardware that speeds up FP8 and FP4: Yes
Compute cores: 5,888

Multi-GPU and connectivity

NVSwitch Lets every GPU in a server talk to every other at full speed: No
PCIe generation: 4

Sharing and security

MIG partitioning Splits one GPU into several isolated smaller GPUs: No
Max MIG instances: 0
Confidential computing: No

Hardware

Architecture: Ada Lovelace
Launch year: 2023
Process node: TSMC 4N
Transistors: 35.8 billion
Die size: 294.5 mm²
Power (TDP) Maximum power draw: 200 W
Cooling: active

Video

NVENC engines: 1
NVDEC engines: 1
AV1 encode: Yes
AV1 decode: Yes

Sources (39)