Publish Date : 2020/01/21

Leadtek Releases NVIDIA Quadro RTX Passive, Professional Graphics Cards with Passive Cooling Solutions


Leadtek, Asia-Pacific channel partner of NVIDIA professional graphics card, releases the passively-cooled NVIDIA Quadro RTX 8000 Passive and RTX 6000 Passive, which declares that NVIDIA Quadro has officially entered the data center from the professional graphics field. Leadtek has complete active and passive cooling solutions for a variety of servers, providing unparalleled visual computing capabilities and scalability for data centers to solve graphics-intensive mixed workloads, complex designs, realistic rendering and driving immersive AR / VR environment applications on 5G networks. Coupled with NVIDIA virtual GPU software, it can accelerate multiple data center workloads.

As the only ultra-high-end RTX GPU with passive cooling and without video connectors, Leadtek NVIDIA Quadro RTX 6000 Passive and RTX 8000 Passive are designed for data centers. Both professional graphics cards are built on the NVIDIA Turing architecture and NVIDIA RTX platform, with 4608 CUDA cores, 576 Tensor cores, and 72 RT cores. The incredible inference performance is ideal for edge computing and data center deployment, as well as AI, real-time ray tracing and high-end graphics processing functions. The major difference is that the Quadro RTX 6000 Passive is equipped with 24GB GDDR6 memory, while the RTX 8000 Passive has 48 GB GDDR6. If connected via NVIDIA NVLink, the RTX 8000 Passive can provide whooping 96GB of ultra-fast GDDR6 memory, significantly boosting AI and data science operations.

In AI and data science, the enhanced Tensor core provides 119.4 TFLOPS Tensor deep learning performance. In addition to the 238.9 TOPS INT8 performance, it also has 14.9 TFLOPS FP32 single-precision and 29.9 TFLOPS FP16 half-precision matrix computing performance, allowing high workloads to be executed more efficiently and reducing data center construction costs.

With RTX GPU support, users can use NVIDIA CUDA X-AI and NGC containers to obtain GPU-accelerated libraries such as cuDNN, cuBLAS, and TensorRT for higher performance for deep learning inference and high-speed computing (HPC) applications. The latest NVIDIA NGX software development tools provide a pre-trained network of deep learning technology to help developers accelerate and optimize image integration applications. For example, the Image InPainting model provides an artificial intelligence imaging technology that allows users to cut out unwanted content, and replace the cutouts with extremely realistic computer-generated background pictures.

Most importantly, the entire Quadro family offers the most stable drivers and ISV certifications for over 100 professional applications and IT management tools. In addition, Leadtek provides an exclusive 3-year warranty and a global technical service center to allow users to enjoy the most professional and secure services.