L4

Data Center GPU

Ada Lovelace NVIDIA L-Series Data Center Server GDDR6 2023 NVIDIA

Save this GPU, then open another GPU page to compare them side by side.

At a Glance

Entry Type: Data Center GPU
Architecture: Ada Lovelace
Series: NVIDIA L-Series
GPU Die: AD104
Launch Date: 2023-03-21
CUDA Cores: 7,424
TMUs: 240
ROPs: 80
Memory: 24 GB GDDR6
Bus Width: 192-bit
TDP/TBP: 72 W
Rec. PSU: 250 W
Interface: PCIe 4.0 x16
Data Confidence: Partial

GPU Specs

Identity
Entry Type	Data Center GPU
Official Name	NVIDIA L4
GPU Codename	AD104
Architecture	Ada Lovelace
Launch Date	2023-03-21
Market Segment Notes	Data Center in NVIDIA L-Series
Core Specs
GPU Die	AD104
Process Node	TSMC 5 nm
Transistor Count	35.8 billion
Die Size	294 mm²
CUDA Cores	7424
SM Count	60
TMUs	240
ROPs	80
AI Units / Tensor Cores	240
RT Units / RT Cores	60
Base Clock	795 MHz
Boost Clock	2,040 MHz
Pixel Fillrate	163.2 GP/s
Texture Fillrate	489.6 GT/s
FP32 Throughput	30.29 TFLOPS
FP16 Throughput	30.29 TFLOPS
FP64 Throughput	0.47 TFLOPS
Memory
Memory Size	24 GB
Memory Type	GDDR6
Memory Bus Width	192-bit
Memory Clock	1,563 MHz
Memory Bandwidth	300.1 GB/s
L2 Cache	48 MB
L1 Cache	128 KB
Board and Display
TDP / TBP	72 W
Recommended PSU	250 W
Power Connectors	None
Interface	PCIe 4.0 x16
Slot Width	Single-slot, 169 mm length
Card Length	169 mm
Board Partner Model	Reference or OEM specification
Reference Design	NVIDIA reference design
Software
CUDA Compute Capability	8.9
Shader Model	6.8
DirectX Support	DirectX 12.2
Vulkan Support	Vulkan 1.4
OpenGL Support	OpenGL 4.6
OpenCL Support	OpenCL 3.0
Driver Branch Notes	Driver support depends on architecture, operating system, and current NVIDIA driver branch.
Windows Support Notes	Check the current NVIDIA driver branch and OEM support policy for this exact GPU.
Linux Support Notes	Check the current NVIDIA Linux driver branch, distribution packages, and kernel compatibility for this exact GPU.

Overview

L4 is an NVIDIA GPU entry in the NVIDIA L-Series family. It is listed with Ada Lovelace architecture and AD104 silicon.

Key reference data: 7,424 CUDA cores/shaders, 24 GB GDDR6 memory, 192-bit memory bus, 72 W board power class.

Server accelerators should be checked against platform compatibility, cooling, power delivery, firmware, and driver or CUDA support.

Why It Mattered

L4 is a NVIDIA L-Series accelerator from the Ada Lovelace generation. Its historical context is data-center deployment rather than gaming, so form factor, memory capacity, interconnect, virtualization features, and server compatibility are the details that make this entry useful.

Collector Notes

For archival use, record the accelerator form factor, memory capacity, board part number, firmware or VBIOS, cooling type, bracket, interconnect support, power connectors, and the server or workstation it came from.

Used-Market Caution

Before buying used, confirm PCIe versus SXM or OAM form factor, passive airflow needs, auxiliary power, HBM size, firmware locks, ECC or retired-page errors, seller proof from nvidia-smi under load, and whether any vGPU or enterprise licensing is required.

Known Issues

Known issues pending source review.

SourcesPartial

Sources reviewed: 2026-06-20

This entry is still being source-checked.

Specifications are currently based on reference/spec-dump sources. Historical notes require additional source review.

Data confidence and source status

Partial entry. Core specifications may be present, but history, variant notes, known issues, and source review may be incomplete.

Verified: core specifications and historical notes have been reviewed against listed sources.
Partial: core specifications may be present, but history, variant notes, known issues, and source review may be incomplete.
Imported: spec-only imported entry pending editorial source review.
Needs source review: narrative notes are incomplete until reviewed.

Independent reference sources

TechPowerUp GPU SpecsModel-level specification reference for this GPU entryOpen source

Derived datasets and spec dumps

RightNow GPU DatabaseOpen data inventory used for bulk normalization and cross-checking Derived/spec-dump source. Treat as a reference dataset, not independent historical verification.Open source

Found an incorrect spec, missing source, or variant issue?

Submit a correction