NVIDIA L4
資料中心A power-efficient, PCIe-form-factor data center GPU based on the Ada Lovelace architecture, optimized for AI inference and graphics workloads.
主要規格
VRAM
24 GB GDDR6
記憶體匯流排
256-bit
FP16
30.29 TFLOPS
FP32
30.29 TFLOPS
TDP
72 W
最佳使用案例
AI Inference
Delivers high-performance tensor core acceleration for running large language models and generative AI applications at scale with low latency.
Virtual Workstations
Enables immersive cloud-based graphics, CAD, and video editing experiences through hardware-accelerated encoding and decoding capabilities.
GPU Cloud Services
Ideal for service providers offering cost-effective, high-density GPU compute instances in the cloud due to its low power consumption per core.