As a long-term partner of NVIDIA, Leadtek Technology focuses on the sales and promotion of products in professional fields such as professional graphics cards, AI, and HPC. In the NVIDIA Ampere generation, NVIDIA RTX A series professional graphics cards were launched to meet the stable computing performance required for professional work such as digital content creation, 3D rendering, ray tracing, AI training/inference, etc. This time, the LEADTEK NVIDIA RTX A4000 professional graphics card is used. Come and try digital content, rendering creative performance.
Specifications
GPU Architecture: NVIDIA Ampere
CUDA Parallel Processing Cores: 6144
NVIDIA Tensor Cores: 192
NVIDIA RT Cores: 48
GPU Memory: 16GB GDDR6 ECC
Single Precision Performance: 19.2 TFLOPS
RT Core Performance: 37.4 TFLOPS
Tensor Performance: 153.4TFLOPS2
Memory Interface: 256-bit
Memory Bandwidth: 448 GB/s
Maximum Power Consumption: 140 W
Graphics Bus: PCI Express 4.0 x16
Display Connector: DP 1.4 x4
Dimensions: 4.4” H x 9.5” L Single Slot
Weight: 500g
External Power Supply Connector: 1x 6-pin PCIe
Sync Output: Support Quadro Sync II
Codec Engine: 1x encode, 1x decode (+AV1 decode)
Graphics API: DirectX 12 Ultimate, Shader Model 6.6, OpenGL 4.6, Vulkan 1.3
Computation API: CUDA 11.6 , DirectCompute, OpenCL 3.0
NVIDIA RTX Vs GeForce Who Is Better For Professional Creation?
How to choose between NVIDIA RTX professional graphics card and NVIDIA GeForce game graphics card for professional video, graphics, computing, rendering and AI workloads?
In fact, this can be explained from “hardware”, “purchase method” and “service”. Taking the NVIDIA RTX A4000 as an example, it adopts the Ampere micro-architecture and has 6144 GPU CUDA cores. This specification is similar to the GeForce RTX 3070 Ti, but there are still quite a few functional differences between the two.
The NVIDIA RTX A4000 is equipped with a larger-capacity GPU memory, with 16GB GDDR6 memory capacity, and supports the error correction ECC function on the memory to ensure accurate long-term computing and rendering results. In addition, it adopts a standard length of 24cm, full height, and a single-slot design. It also supports the design of Quadro Sync II, which is more suitable for the use of multiple GPUs in the professional field.
NVIDIA professional graphics cards are mainly sold by partners, such as the professional graphics card products sold by Leadtek Technology, which enjoy a “three-year warranty”, and “delivery to the house” service is provided within the warranty, and “0800-600″ -206” technical service line, which can provide enterprises and customers with complete technical support and services; this is also the biggest difference from game graphics cards.
At the same time, NVIDIA provides driver branches for “official branch/Studio” and “new function branch” for professional graphics cards. In short, “official branch/Studio” provides the best stability and performance, as well as the most stable version with ISV certification, long life cycle, security updates, etc.; while “new feature branch” provides the latest driver functions, bugfixes, and operating system support.
After all, professional users do not need intensive update functions or support for new hardware as long as the development environment is set up, so the driver will be based on the most stable and compatible version, unlike the driver of the game graphics card. Frequent driver updates to provide 0-Day support after new games are released.
LEADTEK NVIDIA RTX A4000 Professional Graphics Card Out of the Box
NVIDIA professional graphics card can give users more powerful hardware, performance and complete technical support. The new generation of NVIDIA RTX A4000 professional graphics card adopts Ampere micro-architecture, with 6144 CUDA parallel processing cores, 192 Tensor cores and 48 RTs core, and uses PCIe 4.0 x16 to connect to the CPU.
Professional graphics card with large-capacity 16GB GDDR6 ECC memory, even for rendering, computing and simulation of large projects, there is enough memory to complete the task; while RTX A4000 only needs 1 PCIe 6-pin power supply, the maximum power consumption It is a 140W standard-length, full-height, single-slot professional graphics card.
The NVIDIA RTX A4000 professional graphics card is a 24cm long single-slot interface card design. It adopts a single blower fan for heat dissipation but maintains a short PCB design so that the fan can suck the wind from the front and push it out from the rear to drive the cooling airflow inside the computer. Too much air will squeeze into the radiator inside the graphics card to dissipate heat from the GPU.
The RTX A4000 professional graphics card provides 4 DisplayPort 1.4 ports, and there are DP 1.4 to HDMI connectors available in the accessories.
The RTX A4000 professional graphics card only needs to loosen the screws on the front to remove the metal plate on the surface. This simple design is more conducive to “dust cleaning”. After all, computers such as workstations will accumulate dust over time under long-term computing. There is no need to completely disassemble the graphics card, just open the metal cover on the surface to clean up the dust accumulated inside, which is more convenient for logistical maintenance.
LEADTEK NVIDIA RTX A4000 creation and rendering performance test
In terms of performance evaluation, tools such as SPECviewperf, Blender, Vray, REDSHIFT and OCTANEBENCH, which are referenced in the professional field, are used to test the rendering and computing performance of the NVIDIA RTX A4000 professional graphics card. Two sets of editing software that must be familiar to video creators.
Test Platform
Processor: Intel Core i9-12900K
Motherboard: ROG MAXIMUS Z690 HERO
Memory: CORSAIR DOMINATOR PLATINUM RGB DDR5 32GBx2 5200MHz
Graphics Card: LEADTEK NVIDIA RTX A4000
Power Supply: be quiet! DARK POWER 12 850W
Operating System: Windows 11 Pro 21H2
GPU-Z examines NVIDIA RTX A4000 information, using 8nm process, GA104 GPU with 6144 rendering CUDA cores, and 16384 MB GDDR6 (SAMSUNG) memory, while the GPU default 735 MHz, Boost clock 1560 MHz.
SPECviewperf 2020 is based on the standard graphics performance testing tool developed by professional applications to test various professional computer graphics software such as: 3ds Max, Catia, Creo, Energy, Maya, Medical, SNX, SolidWorks and other graphics testing and engineering simulation.
The test is 1920 x 1080 resolution, and the result is FPS. It can be seen that the RTX A4000 can have a very high FPS performance in various tasks, allowing professional workers to have sufficient graphics performance in creation, simulation, and computing.
Blender is a cross-platform, open-source 3D authoring tool that supports various 3D tasks: Modeling, Rigging, Animation, Simulation, Rendering, Compositing, and Motion Tracking. For the test, use Blender Benchmark 3.3.0 to test the rendering work of the Demo project.
Through the Blender Benchmark 3.3.0 test, the RTX A4000 can reach the number of samples per minute of 1584.99 for the monster scene, 939.85 for the junkshop scene, and 814.2 for the classroom scene.
V-Ray Benchmark is developed by Chaos Group. V-Ray is a ray rendering software designed based on the laws of physics, and this tool can perform calculation tests on the rendered images of ray tracing for CPU and GPU respectively.
RTX A4000 can achieve 1280 vpaths performance with CUDA computing, and 1629 vrays computing power with RTX computing.
REDSHIFT is a fully GPU-accelerated renderer, designed to meet the needs of modern advanced CG images, allowing creative individuals to customize the quality of individual technologies to achieve the best balance between performance and quality, while also supporting major mainstream Powerful renderer for 3D applications.
The RTX A4000 completed the rendering of the image in just 4 minutes and 19 seconds through the Redshift Benchmark test.
OctaneBench is a fast and independent rendering test tool. Its core is OTOY Octane Render, which can make full use of RTX and CUDA technology, and supports mainstream 3D applications, allowing creators to have fast and high-quality rendering output.
The test includes 4 scenarios and cores such as Info channels, direct lighting, and path tracing. The RTX A4000 can provide quite good computing speed and meet the fast rendering capabilities required by creators.
Adobe Premiere Pro 2022 video editing software, using the self-developed Mercury Playback Engine GPU acceleration, and can use NVIDA NVENC encoding to accelerate the image output speed, the test uses the BigMix project, which uses 3 FinalAdjusted_MPE 1920×1080 images to form a 4K timeline for H.264, HEVC format output.
The usage rate of the RTX A4000 is much higher than that of the CPU when outputting video and audio. First, it uses NVIDIA H.264 encoding, which can reach 35.9 FPS when outputting 4K videos, and it only takes 131.7 seconds to complete the work; switching to NVIDIA H.EVC encoding output can maintain 35.6 FPS , 130.1 seconds to complete the output work.
DaVinci Resolve 18 is a purely GPU-accelerated video editing program, including powerful color correction and special effects functions, and directly using CUDA core computing, so that the playback and output of video clips have very good performance.
Outputting images via RTX A4000, Wedding_Heavy_Styles is only 57 seconds, Bride_FaceRefine_Selective Color is only 45 seconds, SuperScale 4x HD_Source takes 2 minutes and 44 seconds, and SuperScale 2x 4KSource takes only 37 seconds to output.
Summarize
LEADTEK NVIDIA RTX A4000 professional graphics card can meet the functions of 3D creation, rendering, computer graphics simulation and calculation required by professional users, as well as editing, playback and output of video creation, and can also promote AI development training and inference. hardware performance.
If this article is helpful for you, please share this article with your friends on social media. Thank you!!
This article is based on the personality of the reviews. You are responsible for fact-checking if the contents are not facts or accurate.
Title: LEADTEK NVIDIA RTX A4000 professional graphics card out-of-the-box test/accelerated AI, 3D rendering, image creation
English 
Chinese (Taiwan) 
Chinese (China) 
Japanese 
Korean