Product Introduction

NVIDIA QuadroRTX 6000

Price: telephone consultation

· Turing GPU

· 4,608 NVIDIA ®  CUDA ® Core

· 576 NVIDIA ® Tensor core

· 72 NVIDIA ® RT core

·24GB GDDR6 memory

·Up to 624GB/s video memory bandwidth

· 84T RTX-OPS

·10 Giga Rays/s Raycasting

·16.3 TFLOPS FP32 performance

·32.6 TFLOPS FP16 performance

·261 TOPS INT8 performance

·130.5 TFLOPS Sensor Operation

·*High power consumption: 295W

· 4x DisplayPort 1.4

· 1x VirtualLink

Features and Introduction

**Raytracing GPU

NVIDIA® Quadro RTX ™ 6000 Using NVIDIA Turing ™ Architecture and NVIDIA RTX ™ The platform has brought significant progress in the field of computer graphics for more than ten years for professional workflow. Designers and artists can now take advantage of the hardware's powerful features of accelerated ray tracing, deep learning, and * shading to dramatically increase productivity and create outstanding content at an unprecedented speed.

new RT core and Tensor core provide millions of design and creative professionals with powerful functions of real-time ray tracing and artificial intelligence to enhance workflow. RTX 6000 combines NVIDIA NVLink ™ Technology to expand drawing memory and performance to drive * demanding imaging, artificial intelligence, and visual computing workloads. New VirtualLink ® It provides a connection to the next generation of high-resolution VR HMD, allowing you to view your works in an irresistible virtual environment. Welcome to the future of professional visual computing.

For professionals

1. NVIDIA NVLink ™

Connect two by high-speed interconnection GPU, which expands the memory capacity to 48 GB and provides higher performance at a data transfer rate of 100 GB/s.

2. Next generation memory

Equipped with the industry's first 24 GB super high-speed GDDR6 memory can store complex designs, huge building data sets, 8K movie content, etc.

3. NVIDIA Turing GPU architecture

NVIDIA Turing is equipped with a new RT core for ray tracing, 576 Tensor cores for artificial intelligence, and 4608 CUDA cores for parallel computing. It is an advanced graphics processor.

4. VirtualLink ™

Industry first VirtualLink is used to simplify the connection of existing and next-generation high-resolution VR HMDs.

Performance characteristics

*Real time ray tracing acceleration of sex

Turing, combined with a new hardware ray tracing engine, is the industry's * graphics processor that supports real-time ray tracing. A single Quadro RTX 6000 can produce complex professional models, including * physical shadows, reflections, and refractions, so that users can view them immediately. The Quadro RTX 6000 system, working with applications that utilize APIs such as NVIDIA OptiX, Microsoft DXR, and Vulcan ray tracing, will power truly interactive design workflows and provide immediate feedback for unprecedented levels of productivity.

Reinforced Tensor core

Turing GPU is designed for deep learning of matrix operation, which is the core of neural network training and reasoning functions, including the enhanced Tensor core. In addition to INT8 and two new INT4 and INT1 precision modes, it can also accelerate FP16/FP32 matrix operation. Independent floating point and integer data paths can make the workload run more efficiently by using the hybrid mode of operation and address calculation.

Advanced coloring technology

Mesh shading provides a geometric pipeline based on operation to accelerate the processing and elimination of complex geometric models and scenes, and the efficiency can be improved up to 2 times. Dynamically adjusting shading provides more subtle control over GPU computing power distribution (that is, more cycles are used in areas with more details in the scene, and less cycles are used in areas with less details), which can improve performance under the same picture quality, or save half the time to generate shaded pixels and provide similar picture quality. Material space shading is especially suitable for VR experience. You can process shading and geometric samples at different rates, which is more efficient.

Coding /Decoding engine enhancement

H. 264 encoding bit rate/quality improvement to adjust the performance to reduce the bit rate (for example, about 6% at the same performance level, and about 11% at the same performance level). The drawing memory output of the encoded bit stream and dynamic vector can speed up the post processing of the end-to-end GPU workflow. 4x4 granularity optical flow to improve image grid interpolation/extrapolation, object tracking, motion segmentation, motion recognition, etc. I-frames are only used for H.264 and HEVC decoding, and can only be used for reasoning with I-frames to speed up image monitoring and other purposes. VP9 10/12 bit decoding supports VP9 HBD content decoding.

high speed GDDR6 memory

The Quadro RTX 6000 is equipped with 24GB * advanced GDDR6 memory, and its throughput is 55% higher than that of the previous generation GDDR5x technology. It is extremely suitable for applications that are specialized in processing large datasets and are sensitive to latency.

NVIDIA GPU BOOST 4.0

Automatically * maximize application performance without exceeding the card's power consumption and cooling range. It allows the application to stay in the acceleration pulse state for a longer time at higher temperatures before it falls back to the basic pulse set at the second temperature. This function requires a software application to start, not a separate program. For availability details, please contact   quadrohelp@nvidia.com 。  

Error correction code in drawing memory  (ECC)

Meet the strict requirements of mission critical applications for data integrity, and provide incomparable computing reliability and reliability for workstations.

Advanced Streaming Multiprocessor (SM) Architecture

The Turing architecture is based on the Volta GPU, which is greatly enhanced by SM. Its energy efficiency is 50% higher than that of the previous generation Pascal, and the efficiency is greatly improved within the same power consumption range.

Graphics preemption

Pixel level preemption provides more subtle control and better support for time related work, such as VR dynamic tracking.

Calculate preemption

Instruction level preemption provides more granular control over computing to avoid long running applications monopolizing system resources or timeouts.

many GPU technology

NVIDIA® NVLink

use NVLink connects two Quadro RTX 5000 to double the available memory, and can expand the application performance by using the GPU to GPU data transmission rate of up to 25GB/s (bidirectional), with a total bandwidth of up to 50GB/s.

NVIDIA® SLI ® Technology

Leverage multiple GPU dynamically expands drawing efficiency, enhances image quality, expands display space, and accelerates VR imaging.

utilize NVLink technology extends application performance

image quality

Panoramic anti aliasing  (FSAA)

Use up to 64X FSAA (128 times in SLI mode) significantly reduces visual aliasing artifacts or "aliasing" to obtain * image quality and extremely realistic scenes.

32K material and imaging processing

Material from and imaged to 32K x 32K surfaces to support image processing applications that require * high resolution and quality.

Display Properties

VirtualLink™

Based on open industry standards, VirtualLink can double the resolution of existing VR helmets to support a more immersive experience of the next generation of HMDs. VirtualLink will display that multiple flat cables such as USB and power supply are integrated into a single connection line, reducing the complexity of setting.

DisplayPort 1.4

Support * four more 5K screen @ 60Hz, or two 8K monitors per card. The Quadro RTX6000 supports HDR colors, including 4K @ 120Hz 10/12b HEVC decoding and up to 4K @ 60Hz 10b HEVC encoding. Each DisplayPort connector can drive 4096x2160 @ 120 Hz ultra-high resolution and 30-bit color.

NVIDIA® nView ® Advanced desktop software

In a single large display or multi display environment, end users can gain unprecedented control over desktop experience to improve productivity.

NVIDIA® Mosaic ™ technology

Extends desktops and applications from a single workstation to * multiple 4 GPUs and 16 monitors, providing complete performance and image quality.

NVIDIA® Quadro Sync III

In a single system Synchronize among 8 GPUs * Display and image output of more than 32 displays (connected through two Sync II adapter cards), reducing the number of machines needed to establish a high-level image visualization environment.

Frame page lock  (Frame Lock)

Each frame page lock connector is designed with an automatic locking and holding mechanism to ensure the connection with the frame page lock flat cable, providing strong connectivity and high productivity.

OpenGL four buffer stereo support

Provide professional applications with smooth and experienced 3D experience.

Support ultra-high resolution desktop

At * large Get more Mosaic topology options on a 32K desktop size high-resolution display.

major 3D stereo synchronization

Stable control of 3D effect, directly connect 3D hardware is synchronized to the Quadro graphics card.

Software support

in the light of Turing optimized software

Deep learning frameworks such as Caffe2, MXNet, CNTK, TensorFlow, etc. can significantly speed up training time and improve multi node training efficiency. GPU acceleration libraries such as cuDNN, cuBLAS, and TensorRT provide higher performance for deep learning reasoning and high-speed computing (HPC) applications.

NVIDIA® CUDA ® Parallel computing platform

Native implementation standard program languages such as C/C++and Fortran, as well as APIs such as OpenCL, OpenACC and Direct Compute, to speed up ray tracing, film and image processing, and fluid dynamics computing.

Single memory

Single seamless The 49 bit virtual address space allows data to move transparently within the memory fully allocated by the CPU and GPU.

NVIDIA® GPUDirect for Video

GPUDirect for Video accelerates communication between GPU and video I/O devices by avoiding unnecessary duplication of system memory data and CPU burden.

NVIDIA Enterprise Management Tools

Maximize the system's uptime, seamlessly manage large-scale deployment and remotely control graphics and display settings to achieve efficient operation.

Specifications

<tr style="box-siz