GPU Wiki

Compare GPUs

Select up to four GPU Wiki entries to compare matching specifications row by row. The comparison uses the site database and keeps generation-specific specs visibly separate.

Popular comparisons

Add a third or fourth GPU
Spec L40S
Overview
Product name NVIDIA L40S
Architecture Ada Lovelace
Generation / family NVIDIA L-Series
GPU codename AD102
Launch date 2023-08-08
Market segment Data Center in NVIDIA L-Series
Board type / entry type Data Center GPU
GPU
Process node TSMC 5 nm
Die size 609 mm²
Transistor count 76.3 billion
GPU chip AD102
SM / multiprocessor count 142
CUDA Cores 18176
Tensor Cores 568
RT Cores 142
ROPs 192
TMUs 568
Clocks
Base / Core clock 1,110 MHz
Boost clock 2,520 MHz
Memory clock 2,250 MHz
Memory
Memory size 48 GB
Memory type GDDR6
Memory bus width 384-bit
Memory bandwidth 864 GB/s
Bus / Interface
Bus interface PCIe 4.0 x16
AGP / PCI / PCIe distinction PCIe
Power
TDP / TBP / board power 350 W
Recommended PSU 700 W
Power connectors 1x 16-pin
Slot width Dual-slot, 267 mm length
Display / Media
Display outputs 1x HDMI 2.1, 3x DisplayPort 1.4a
Features
DirectX support DirectX 12.2
OpenGL support OpenGL 4.6
Vulkan support Vulkan 1.4
CUDA support 8.9
Shader Model 6.8
Ray tracing support Hardware RT units present
Performance / Benchmarks
FP32 throughput 91.61 TFLOPS
FP16 throughput 91.61 TFLOPS
Muted cells indicate unavailable, not applicable, or not directly comparable specifications.

Frame rates compare runs of the same game at the same quality preset on one test system. DLAA counts as native resolution, because it applies anti-aliasing at full render resolution rather than upscaling; DLSS and FSR upscaling are never counted as native. Where a capture named ray tracing without a readable on or off value, it is read as on.