NVIDIA's full name is "NVIDIA RTX 4000 SFF Ada Generation". The datasheet explicitly says vGPU software support: No, and NVIDIA NVLink: No, so neither is recorded on this row. Its 306.8 TFLOPS tensor figure is footnoted as effective FP8 using the sparsity feature, giving 153.4 dense.

NVIDIA RTX 4000 SFF Ada 20GB

Ada Lovelace PCIe 2023 4nm Spec confidence: Official
FP8 (dense)
153
TFLOPS
FP32
19
TFLOPS
VRAM
20 GB
GDDR6
Bandwidth
280 GB/s
memory
TDP
70 W
274.3 TFLOPS/kW

Overview

The NVIDIA RTX 4000 SFF Ada Generation fits 6,144 CUDA cores and 20 GB of GDDR6 with ECC into a half-height, 70 W card that draws all its power from the slot. It is the same silicon configuration as the full-height RTX 4000 Ada but clocked to suit the envelope, giving 19.2 TFLOPS of FP32 and 153.4 dense TFLOPS of FP8 with 280 GB/s of memory bandwidth. For inference deployments where space and power are the binding constraints, small form factor desktops, 1U servers, industrial and medical systems, it delivers considerably more memory than any other card at this power level from its generation. Four Mini DisplayPort 1.4a connectors handle display output. Its Blackwell successor is the RTX PRO 4000 Blackwell SFF Edition, which lifts the same 70 W envelope to 24 GB and 24 TFLOPS.

Performance

Peak theoretical throughput by precision type

PrecisionDense2:4 Sparse
FP64
64-bit floating point
0.3TFLOPS Structured sparsity is a tensor-core feature; this vector precision has no sparse form
FP32
32-bit floating point
19TFLOPS Structured sparsity is a tensor-core feature; this vector precision has no sparse form
TF32
No verified data available The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision
BF16
No verified data available The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision
FP16
No verified data available The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision
FP8
8-bit floating point
153TFLOPS 307TFLOPS
FP6
No verified data available The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision
FP4
No verified data available The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision
INT8
No verified data available The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision

Specifications

Architecture

Ada Lovelace

Form Factor

PCIe

Launch Year

2023

Process Node

4nm

Memory

20 GB GDDR6

Bandwidth

280 GB/s

TDP

70 W

Max Power

70 W

CUDA Cores

6,144

Spec Confidence

Official

Full Specifications

Compute Engine
CUDA Cores 6,144
Tensor Cores 192 (4th Gen)
Streaming Multiprocessors 48
Memory
VRAM 20 GB
Memory Type GDDR6
Bandwidth 280 GB/s
Interface Width 160-bit
Interconnect & I/O
PCIe 4.0 x16
Power & Thermal
TDP 70 W
Max Board Power 70 W
Cooling Active
Enterprise Features
ECC Memory Yes
Sparsity Yes
Compute APIs CUDA 11.6, OpenCL 3.0, DirectCompute
Physical & Media
Card Length 167.6 mm
Width 2-slot
NVENC Engines 2
NVDEC Engines 2
General
Form Factor PCIe
Architecture Ada Lovelace
Process Node 4nm
Launch Year 2023

Datasheet & Resources

Data Provenance

Every figure traced to a source

Primary Source

Publisher
NVIDIA
Published
2023-04-01

Data Quality

Spec confidence
Official
Clock basis
Boost
Core precisions with figures
3 of 9
Normalization
All values in TFLOPS

Rent this GPU

$0.18 / GPU-hr

Lowest of 2 live listings across 1 provider

See all rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

Similar GPUs

NVIDIA L40S 48GB

PCIe · 2023
FP32: 92 TFLOPS
Compare vs L40S

NVIDIA L40 48GB

PCIe · 2023
FP32: 91 TFLOPS
Compare vs L40

Frequently Asked Questions

How many TFLOPS does the NVIDIA RTX 4000 SFF Ada have?

The NVIDIA RTX 4000 SFF Ada delivers 19 TFLOPS FP32 and 153 TFLOPS FP8 at peak. Flopper does not currently have a verified FP16 throughput figure for it.

What is the power consumption of the NVIDIA RTX 4000 SFF Ada?

The NVIDIA RTX 4000 SFF Ada has a TDP (Thermal Design Power) rating of 70 watts.

How much memory does the NVIDIA RTX 4000 SFF Ada have?

The NVIDIA RTX 4000 SFF Ada is equipped with 20 GB of VRAM with 280 GB/s of memory bandwidth.

What architecture is the NVIDIA RTX 4000 SFF Ada based on?

The NVIDIA RTX 4000 SFF Ada is based on the Ada Lovelace architecture, launched in 2023.

Stay Updated on GPU Releases

Get notified when new GPUs are added or specifications are updated.

Loading verification...

No spam, unsubscribe anytime.

Back to GPUs