for ThinkAgile MX3530-H Hybrid Appliance; MX3531-H Hybrid Certified Node
Overview
The NVIDIA Tesla T4 GPU is a single-slot, low-profile PCIe 3.0 universal deep learning accelerator. It provides multi-precision performance to accelerate deep learning and machine learning training and inference, video transcoding and virtual desktops.
70 W power consumption for energy-efficient operation
320 GBps bandwidth with GDDR6 SDRAM for fast data processing
Fanless design for a quiet and dust-free environment
Supports DirectCompute, OpenACC, and OpenCL APIs for versatile computing
2560 CUDA cores and NVIDIA Turing architecture for significant computing performance
Efficient power consumptionWith an operational power consumption of just 70 W, this GPU computing processor delivers exceptional performance without the high energy costs, making it an ideal choice for energy-conscious setups.Advanced memory technologyEquipped with GDDR6 SDRAM and a bandwidth of 320 GBps, this device ensures rapid data transfer rates and smooth processing for complex computations and high-quality video rendering.Comprehensive video supportSupporting a wide range of APIs including DirectCompute, OpenACC, and OpenCL, alongside 2560 CUDA cores, this processor is designed to meet diverse computing needs.Innovative cooling designThe fanless heatsink design not only minimizes noise but also reduces dust accumulation, ensuring a longer lifespan for the GPU while maintaining optimal performance.Robust architectureLeveraging NVIDIA Turing GPU architecture and equipped with 320 NVIDIA Tensor Cores, it delivers significant performance for computational tasks.