Infrastructure atlas
NVIDIA/Ada Lovelace

L4 Tensor Core

A compact, low-power PCIe accelerator for video pipelines, edge inference, and cost-sensitive production serving.

GPUAdainferencevideoedge
L4 Tensor Core hardware

01 / Memory

24 GB GDDR6

ECC enabled

02 / Bandwidth

300 GB/s

Memory bandwidth

03 / Power

72 W

Low-profile class

Precision & peak compute

PrecisionPublished peakBasis
FP3230.3 TFLOPSDense
FP8 Tensor242 TFLOPSDense
FP8 Tensor485 TFLOPSSparse
INT8 Tensor242 TOPSDense
INT8 Tensor485 TOPSSparse

Peak FLOPS and TOPS are theoretical hardware figures. Sparse and dense modes, chip and system totals, and form factors are not interchangeable.

NVIDIA L4 specifications

Technical specification

Silicon & memory
Architecture
Ada Lovelace
FP32
30.3 TFLOPS
FP8 Tensor
242 TFLOPS dense · 485 sparse
Video
AV1 encode/decode support
System & software
Form factor
Single-slot PCIe
Interconnect
PCIe Gen4
Best fit
Video, speech, edge, efficient inference
Software
CUDA, TensorRT, CV-CUDA