L40S
NVIDIA L40S Full Specs
18,176 Shaders 2.52GHz |
48GB GDDR6864GB/s |
·· 91.61 TFLOPS |
Form Factor PCIe Card |
TDP 300W |
Power Connectors 1x 16-Pin 12VHPWR |
L40S2.93 POPSINT4 Tensor Sparse3.35 PFLOPS FP4 Tensor SparseGeForce RTX 5090
x1.14Clock Speed ··· |
Peak OPS 2.93 POPSINT4 Tensor Sparse |
Tensor FP8-16 732.9 TFLOPSFP8-16 Tensor Sparse 1.47 PFLOPSTensor FP8-32 366.4 TFLOPSFP8-32 Tensor Sparse 732.9 TFLOPS |
Tensor FP16-16 366.4 TFLOPSFP16-16 Tensor Sparse 732.9 TFLOPSTensor FP16-32 183.2 TFLOPSFP16-32 Tensor Sparse 366.4 TFLOPS |
BF16 91.61 TFLOPSTensor BF16 183.2 TFLOPSBF16 Tensor Sparse 366.4 TFLOPS |
Tensor TF32 91.61 TFLOPS |
FP32 91.61 TFLOPS |
FP64 1.43 TFLOPS |
Tensor INT4 1.47 POPS |
Tensor INT8 732.9 TOPS |
Ray 211.7 TOPS |
Pixel Rate 483.8 GPixel/s |
Texture Rate 1.43 TTexel/s |
Shaders 18,176 Shaders |
TMUs 568 TMUs |
ROPs 192 ROPs |
Tensor Cores 568 T-Cores |
RT Cores 142 RT-Cores |
SMs 142 SMs |
Base Clock 1.11GHz |
Boost Clock 2.52GHz |
L1 64KB/SM Tex |
L2 Cache 96MB shared |
48GB GDDR6 |
Memory Bus 384-bit |
Memory Speed 18GT/s |
Memory Bandwidth 864GB/s |
ECC No |
TDP 300W |
Max Resolution 7680x4320 |
Refresh Rate Calculator 7680x4320 Resolution8K UHD Refresh Rate 60Hz |
Multi-Monitor 4 |
Adaptive Sync G-Sync FreeSync |
DSC Supported |
HDCP HDCP 2.3 |
4x DisplayPort 1.4 |
Shader Model 6.6 |
DirectX DirectX 12Direct3D 12_3 |
OpenGL 4.6OpenCL 3Vulkan 1.3 |
CUDA 8.9PureVideo HD VP12VDPAU Feature Set L |
Encoder 2x NVENC 8 |
Codec AVC (H.264) HEVC (H.265) AV1 |
Decoder NVDEC 5 |
Codec MPEG-1 MPEG-2 MPEG-4 VC-1 VP8 VP9 AVC (H.264) HEVC (H.265) AV1 |
Form Factor PCIe Card |
PCIe PCIe 4.0 x16 2-Slots |
Height 111mm (4.37")Width 267mm (10.51")Depth 40mm (1.57") |
Cooling Passive |
Power Connectors 1x 16-Pin 12VHPWR |
Manufacturer |
Chip Designer |
Architecture |
Family |
Branding  |
Codename NV182 |
Chip Variant AD102-200-A1 |
Market Segment Server |
Release Date Oct 13, 2022 |
Foundry TSMC |
Fabrication Node 4N |
Die Size 608mm² |
Transistor Count 76.3B |
Transistor Density 125.4 MTr/mm² |