NVIDIA L40S Professional Graphics Card
Higher clock and TDP than L40
Product status: Official | Last Update: 2023-08-09 | Report Error
Graphics Processor
AD102-895/895A
Cores : TMUs : ROPs
18176 : 568 : 192
Memory Size
48 GB G6
Memory Bus
384 bit
NVIDIA L40S
Overview
Manufacturer
NVIDIA
Original Series
Data Center Ada
Model
NVIDIA PG133 SKU 250
Graphics Processing Unit
GPU Model
AD102-895/895A (AD102)
Architecture
Ada
Fabrication Process
4 nm (TSMC 4N)
Die Size
608 mm2
Transistors Count
76.3B
CUDAs
18176
Tensor Cores
568
Ray Tracing Cores
142
TMUs
568
ROPs
192
Clocks
Base Clock
TBC MHz
Boost Clock
2520 MHz
Memory Clock
TBC MHz
Effective Memory Clock
TBC Mbps
Memory Configuration
Memory Size
49152 MB
Memory Type
GDDR6
Memory Bus Width
384-bit
Memory Bandwidth
0.0 GB/s
Physical
Interface
PCI-Express 4.0 x16
Height
2-slot
Power Connectors
1× 16-pin
TDP/TBP
350 W
Recommended PSU
700 W
Multi-GPU Support
2-way
Display Outputs
DisplayPort
4 × (DP 1.4a)
API Support
DirectX
12.2
Vulkan
1.2
OpenGL
4.6
OpenCL
3.0
Performance
Pixel Fillrate
483.8 GPixels/s
Texture Fillrate
1.4 TTexel/s
Peak FP32
91.6 TFLOPS
FP32 Perf. per Watt
261.7 GFLOPS/W
FP32 Perf. per mm2
150.7 GFLOPS/mm2
| Model | Cores | Boost Clock | Memory Clock | Memory Config. | |
|---|---|---|---|---|---|
 | NVIDIA L40S | - | |||
| | NVIDIA L40 | - |
| Model | Cores | Boost Clock | Memory Clock | Memory Config. | |
|---|---|---|---|---|---|
| | NVIDIA L40 | - | |||
| | NVIDIA L40S | - | |||
 | NVIDIA RTX 6000 ADA | ||||
 | NVIDIA GeForce RTX 4090 | ||||
| | NVIDIA GeForce RTX 4090D | ||||
| | NVIDIA RTX 5880 ADA | ||||
 | NVIDIA RTX 5000 ADA | ||||
| | NVIDIA GeForce RTX 4070 Ti SUPER (AD102) |