NVidia L40-4Q

Description

nVidia L40-4Q- Graphics Card Specifications

Brand NVIDIA
Model / vGPU Profile NVIDIA L40-4Q
Physical GPU NVIDIA L40
GPU Architecture NVIDIA Ada Lovelace
Memory Size 4 GB framebuffer for L40-4Q vGPU profile
Physical GPU Memory 48 GB GDDR6 with ECC
Memory Type GDDR6 with ECC
Memory Bandwidth 864 GB/s (physical L40)
Memory Clock Speed NVIDIA's official L40 specifications do not list a separate memory-clock speed
TDP / Board Power 300 W maximum (physical L40)
Video Card Category Professional / Data Center Graphics Accelerator
Power Connector 1 × 16-pin PCIe power connector
Recommended PSU NVIDIA specifies a 300 W maximum board power; actual PSU requirement depends on the complete server/workstation configuration
Maximum Resolution Up to 7680 × 4320 (8K) supported by the L40-4Q profile
Memory Interface GDDR6; NVIDIA's published L40 specifications do not separately state a memory-bus-width figure
OpenGL Professional OpenGL support; NVIDIA does not specify an OpenGL version in the L40 product specifications
Cooling Passive cooling
Display Outputs 4 × DisplayPort 1.4a on the physical L40
PCI Express PCI Express Gen4 x16
CUDA Cores 18,176
RT Cores 142, 3rd Generation
Tensor Cores 568, 4th Generation
Form Factor Full-height, full-length, dual-slot
Virtualization NVIDIA vGPU supported
L40-4Q Framebuffer 4,096 MB
Maximum L40-4Q vGPUs per physical GPU Up to 12 in equal-size mode
L40-4Q Intended Use Virtual Workstations