The PNY Tesla T4 professional graphics card accelerates a wide variety of cloud-based workloads such as high performance computing, deep learning training and inference, machine learning, data analysis and graphics processing. With no graphics output, it is intended for server use in data centres. On the technical side, the Tesla T4 is equipped with a Turing GPU for a total of 2,560 Cuda cores, 320 Tensor cores, 16 GB GDDR6 memory with 256-bit interface to offer a computing power of 65 Tflops in mixed precision.
The PNY Nvidia Tesla T4 graphics card delivers up to 40x performance with low latency, enabling more real-time requests to be processed.
NVIDIA Tesla T4 GPU accelerates a wide variety of cloud workloads such as high performance computing, deep learning training and inference, machine learning, data analysis, and graphics processing. Based on the new NVIDIA Turing architecture and built on a compact, low-power PCIe configuration (70 watts), the NVIDIA T4 has been optimised for consumer computing servers and is designed to deliver unprecedented multi-precision computing capabilities with its Turing Tensor cores and new RT cores.
NVIDIA Tesla T4 builds on Tensor's revolutionary Turing technology to deliver multi-precision computing performance with a wide variety of workloads. With unprecedented FP32, FP16, INT8, and INT4 performance, this accelerator delivers up to 40 times higher performance than CPU-based configurations.