The new generation of NVIDIA Ada Lovelace professional graphics cards has arrived. Nvidia RTX™ 5000 Ada Generation (hereinafter referred to as RTX 5000) also has updated CUDA core, third generation RT core and fourth generation tensor core and is configured with a lower tier 250W graphics card power consumption, 32GB GDDR6 ECC memory, 2 sets of codec engines each supporting AV1, supports virtualization, Quadro Sync II, RTX Desktop and enterprise-specific features. Professional users with different computing needs have a suitable professional graphics card upgrade. Professional graphics card sales and support services are provided by NVIDIA’s long-term partner LEADTEK.
Specific
GPU Architecture: NVIDIA Ada Lovelace
CUDA parallel processing cores: 12800
NVIDIA Tensor Core: 400
NVIDIA RT core: 100
Single precision performance: 65.3 TFLOPS
RT Core Performance: 151.0 TFLOPS
Tensor Performance: 1044.4 TFLOPS2
GPU Memory: 32GB GDDR6 ECC
Memory interface: 256 bit
Memory bandwidth: 576 GB/s
Maximum power consumption: 250W
Graphics bus: PCI Express 4.0 x16
Display connector: DP 1.4×4
Dimensions: 4.4″ H x 10.5″ W double wide
External Power Connector: 1x PCIe 12+4 pin
Synchronous Output: Supports Quadro Sync II
Codec engine: 2x encoding, 2x decoding (+ AV1 encoding/decoding)
Same as the AD102 core NVIDIA RTX 5000 Ada Generation professional graphics card
Modern professional work, be it professional work tasks such as image output, artificial intelligence, 3D rendering, VFX, simulation and computing, can use GPU acceleration to achieve faster processing times, and NVIDIA starts with GPU computing for purposes CUDA generics and gradually develops Tensor core and RT core, and many convenient development functions are overlaid on the hardware, enabling 3D rendering to achieve real-time ray tracing, and combining AI computing, robot simulation and OpenUSD to create the Omniverse platform.
The previously tested flagship NVIDIA RTX 6000 Ada Generation (hereinafter referred to as RTX 6000) has a full AD102 core and 48 GB GDDR6 ECC memory. However, the relative cost of a single card and power consumption of 300W mean that typical devices from small to medium-sized enterprises and large laboratories must switch to Ada. The cost is slightly higher for each generation, so NVIDIA has recently updated the new Ada generation cards such as RTX 5000, RTX 4500 and RTX 4000 to meet the upgrade needs of various professionals.
↑ NVIDIA RTX Ada generation.
RTX 5000 also uses the AD102 core and reduces it to 100 sets of SM configurations, has basic specifications such as 12800 CUDA, 400 Tensor and 100 RT, and is equipped with 32GB GDDR6 ECC memory with a bit width of 256 bits and a 576 GB/s memory bandwidth; Additionally, the GPU power consumption is only 250W and requires a 16-pin PCIe power supply or 2 x 8-pin PCIe adapters, allowing most professional workstations to be directly upgraded.
↑ NVIDIA RTX 5000 Ada generation data sheet.
The difference between NVIDIA RTX professional graphics card and GeForce RTX gaming graphics card is that the professional graphics card is equipped with larger GDDR6 video memory, supports memory ECC error correction code function, and has an operating clock more stable and lower performance. Consumption settings.
Additionally, NVIDIA RTX professional graphics cards support multiple enterprise professional applications, such as RTX Virtual Workstation (vWS), 2 sets of 5th Generation NVDEC decoding engines and 8th Generation NVENC encoding engines, Mosaic multi-display, desktop management software RTX, Quadro Sync II, plus business management tools and more.
Additionally, NVIDIA professional graphics cards are primarily sold by partners. For example, professional graphics cards sold by Leadtek Technology have a “three-year warranty.” Within the warranty, “home delivery” service is provided, and “0800-600 -206” technical service hotline can provide enterprises and customers with comprehensive technical support and services; this is also the biggest difference compared to gaming graphics cards.
Welcome to leave your information here to learn more about NVIDIA. Exclusive Agent – Jieyuan will introduce you and serve you later. You can also be the first to receive the latest products and preferred information.
↑ 3-year warranty, door-to-door delivery and free customer service hotline.
At the same time, NVIDIA provides driver branch versions of the “Official/Studio branch” and the “New Features branch” for professional graphics cards. Simply put, the “official branch/studio” provides the best stability and performance, as well as the most stable version with ISV certification, longer lifecycle, security updates, etc.; while the “new features branch” provides the latest driver features, bug fixes, and operating system support.
↑ NVIDIA RTX Driver Download.
Unboxing of NVIDIA RTX™ 5000 Ada Generation Professional Graphics Card / 26.7cm Length, 2 Slots, Workstation Design with Fan
The RTX 5000 professional graphics card adopts a matte black square and touchable appearance with this generation’s classic dark gold embellishment on the curves on the side of the graphics card. The graphics card is 10.5 inches long, has 2-slot installation space, and is a workstation design with a single drum fan for heat dissipation. It is in line with the workstation requirements. Multi-board system expansion with heatsink flow consistency.
↑RTX 5000 external box.
↑ Unboxing.
Unlike the Mirror Black styling of the RTX 6000, the RTX 5000 has an all-matte black look with sharp, square lines, as well as a traditional fan cooling design, which is rare these days. The dark gold frame and The unique curved shape on the side of the graphics card facilitates professional drawing. The card has a unique design sense.
↑ RTX 5000 front.
↑ RTX 5000 back.
There is a Quadro Sync II connection interface in the upper left corner of the graphics card side. On the front of the graphics card is the word NVIDIA and the 12VHPWR 16-pin PCIe power socket. Accessories include one 16-pin PCIe adapter cable to two 8-pin PCIe adapter cables, so workstations and servers can be upgraded without replacing the power supply.
↑Word NVIDIA.
↑ Quadro Sync II connection interface.
↑ PCIe 16 pin 12VHPWR power supply.
The RTX 5000 professional graphics card has 4 DisplayPort 1.4 ports and a DP 1.4 to HDMI connector is available as an accessory. It can output up to four professional 4K@120Hz or 5K@60Hz displays simultaneously. Of course, it can also support two outputs for ultra-high-resolution 8K@60Hz displays.
↑ 4 DisplayPort 1.4 ports.
↑ DP 1.4 to HDMI.
↑ PCIe 16-pin to PCIe 8-pin adapter cable.
NVIDIA Control Center, ECC, Mosaic, RTX Desktop
The RTX 5000 uses 32GB GDDR6 ECC memory, and the memory ECC feature is enabled by default. If you want to make the change, you can find the ECC Edit status in the NVIDIA Control Center and set it to On or Off. After all, calculations and simulations for professional work are time-consuming, so memory error correcting code (ECC) can improve the accuracy of calculations and the reliability of the workstation.
↑ ECC status.
For multi-GPU and multi-display connection scenarios on professional workstations, NVIDIA Control Center also provides topology information, including system driver version, 4 screen output ports of professional graphics cards, and information about the currently connected external display, enabling users to more effectively The management workstation currently uses a GPU with an external monitor.
↑ System topology。
NVIDIA Mosaic multi-monitor technology supports display integration with up to 4 GPUs and 16 monitors. Users can organize monitors according to usage environment and integrate them into a single desktop. It also provides advanced multi-screen features such as screen frame correction and projector overlay Built-in features.
↑ NVIDIA Mosaic。
RTX desktop management software is a more powerful tool for organizing and managing desktop windows. It can control the arrangement of all windows on multiple screens directly in the software and can save the current window configuration as a profile and provide default settings. Different GRID segmentation methods allow you to more effectively handle multi-screen work situations.
↑ RTX desktop management software can organize the window positions of each screen directly in the software.
↑ You can also save your settings as a profile.
↑ And provide a GRID split screen.
In Control Center you can also adjust how the GPU is used. The default mode is used for both drawing and calculations. If you are performing drawing functions, you can also switch to GPU mode which focuses on drawing tasks.
↑ GPU usage patterns.
Testing the creation and rendering performance of the NVIDIA RTX™ 5000 Ada generation
In terms of performance evaluation, tools such as SPECviewperf, Blender, Vray and OCTANEBENCH, often cited in the professional field, are used to test the rendering and computing performance of the RTX 5000 professional graphics card. DaVinci Resolve 18 audio and video are also included the output tests, Omniverse Creator and Stable Diffusion.
test platform
Processor: Intel Core i9-14900K
Motherboard: ASUS ROG MAXIMUS Z790 DARK HERO
Memory: Crucial DDR5-5600 16GBx2
Graphics card: NVIDIA RTX™ 5000 Ada generation
System disk: Solidigm P41 Plus 1TB PCIe 4.0 SSD
Radiator: ROG ROG RYUJIN III 360
Power supply: seasonal VERTEX GX-1000
Operating system: Windows 11 Pro 22H2 64bit, resizable BAR activated
Driver version: NVIDIA 537.99
GPU-Z Checking the information of the RTX 5000, the AD102 GPU using the 5nm process has 12,800 rendering CUDA cores and 32,768 MB of GDDR6 memory (Hynix), and the default GPU clock is 1155 MHz and Boost 2550 MHz.
↑ Z GPU
CINEBENCH 2024 The new version is fully developed using Cinema 4D’s default Redshift rendering engine and includes comprehensive GPU and CPU testing and supports cross-platform performance comparisons, which also results in more powerful rendering tests for the CPU.
The RTX 5000 GPU scored 20137 points, which is much faster than the mainstream i9-14900K processor.
↑ CINEBENCH 2024.
Da Vinci Resolve 18 The video editing program, which is purely GPU accelerated, also includes powerful color correction and special effects functions and directly uses CUDA core operations to achieve excellent performance in video editing playback and output.
↑ DaVinci Resolve 18.
First, the first test project uses Blackmagic RAW 4K images and has a Wedding_Heavy_Styles timeline. This video uses many Resolve effects, such as OFX: Rays of Light / Glow / Sketch, etc., and produces a very high-style video.
Bride_FaceRefine_Selective_Color uses Face Refinement to track faces and color code the bride’s subject; 50% Retime and Optical Flow Enhanced Better both use Optical Flow technology to reduce image speed by 50%.
SuperScale2x 4K Source uses 4K ProRES source video to produce 4K video output with 2x zoom; SuperScale4x HD_Source uses H.264 HD source video and uses Resolve Super Scale to produce 4K video.
RTX 5000 features 2 sets of 8th generation NVENC encoding engines, which have very fast processing capabilities for 4K audio and video output. Even if 4 HD are combined into the 4K output, it only takes 1 minute and 1 second to complete the task, and the remaining outputs are all within 1 Done in minutes.
↑ DaVinci Resolve 18 audio and video output.
The second test is an AV1 and HEVC encoding test with dual NVENC encoding. The test project is a 44-second short film from the Blender Open Movie Project “Tears of Steel” and features Prores442HQ 8K 30FPS and Prores422HQ 4K 30FPS video, available for testing the output performance of HEVC and AV1 encoding.
The output settings mainly use NVIDIA Encoder, Quality: Limit to 80000 Kb/s, Encoding Profile: Main, Speed Control: Constant Bitrate, Preset: Fastest, Optimization: High Quality, Two Pass: Disable and other output settings.
↑ 8K, AV1 test project.
For RTX 5000, it is really very easy to produce 4K projects, and if you upgrade to 8K video output, you can get very fast output performance regardless of HEVC or AV1 encoding.
↑ DaVinci Resolve 18 audio and video output.
Mixer It is a cross-platform and open-source 3D creation tool that supports various 3D operations: modeling, rigging, animation, simulation, rendering, compositing and motion tracking, etc. For testing, Blender Benchmark 3.5.0 was used to test the rendering work of the Demo project.
The Blender Benchmark RTX 5000 test achieved 3146,377 points in the monster scene, 1386,968 points in the junk dealer scene, and 1515,451 points in the classroom scene, which meets the rendering performance required for 3D creation.
↑ Blender Benchmarks。
Benchmark V-Ray Developed by Chaos Group, V-Ray is a light rendering software designed based on physical laws. This tool can perform computational testing of rendered ray tracing images for CPU and GPU respectively.
In the V-Ray test, the RTX 5000 can achieve a computing performance of 2002 vpaths per second using CUDA computing and a computing performance of 2474 vrays per second using RTX computing.
↑ V-Ray CUDA.
↑ V-Ray RTX.
Octane Bench It is a fast and independent rendering test tool, the core of which is OTOY Octane Render, which can take full advantage of RTX and CUDA technology and supports mainstream 3D applications, allowing creators to achieve fast and high-quality rendering output.
The test included 4 scenes and cores such as information channels, direct lighting and path tracking. The RTX 5000 can provide significant processing speed to meet the fast rendering capabilities that creators demand.
↑ Octane bank.
SPECviewperf 2020 Standard drawing performance test tool developed based on professional applications, testing various professional computer graphics software such as: 3ds Max, Catia, Creo, Energy, Maya, Medical, SNX, SolidWorks and other drawing tests and engineering simulations.
The test has a resolution of 3840 x 2160 and the result is FPS. It can be seen that the RTX 5000 can have very high FPS performance in various tasks, allowing professionals to have sufficient graphics performance when creating, simulating and computing.
↑ SPECviewperf 2020.
NVIDIA RTX™ 5000 Ada Generation Omniverse Creator Test
NVIDIA Omniverse The collaboration platform that accelerates 3D design not only seamlessly supports the industry’s leading 3D apps, allowing creators around the world to create 3D designs simultaneously and achieve real-time rendering effects, but also features features such as Digital Twin and robot training simulation.
This test primarily uses the Omniverse Create project to conduct real-time light tracking rendering and DLSS 3 accelerated FPS performance testing. Projects tested include: Campfire_Main_03_Flow3 simulates physical campfire effects, Flowers4 tests instance geometry and animation of particle flow, Incense_Main tests the simulation of smoke and changes in reflective materials, the JadeTiger test scene contains a detailed model of advanced jade materials, and the Warehouse_Physics_Stage test scene is a simulation of the domino effect of collapsing warehouse shelves .
↑ Create the Omniverse。
In the UHD ViewPort and DLSS3 Performance settings, the RTX 5000 averages performance of 42.49 FPS in Campfire, 33.19 FPS in Flowers4, 45.63 FPS in Incense_Main, 45.69 FPS in JadeTiger, and 37.95 FPS in Warehouse_Physics_Stage.
↑ Test of creation of the Omniverse.
NVIDIA RTX™ 5000 Ada Generation AI Graphics Stable Spread Test
Stable Diffusion, the popular AI deep learning text-to-image generation model, was manually installed using the standard version released by Automatic1111, and the expanded and imported TensorRT optimization allows Stable Diffusion to achieve the best GPU acceleration performance.
↑ Stable diffusion.
For test settings and spells, please refer to the description below the image. Through testing, RTX 5000 TensorRT can achieve an image generation speed of 18.46 images per minute. Compared to just 5.77 images per minute on the standard Automatic1111, performance is improved by 220%.
↑ Stable diffusion performance test.
Power consumption and temperature measurements of the NVIDIA RTX™ 5000 Ada generation
For graphics card temperature and power consumption testing, Omniverse and DaVinci Resolve 18 were used to run the project 3 times to record GPU temperature and power consumption. The highest power consumption was tested using Furmark for a burn-in test. When measuring power consumption, use the PACT tool provided by NVIDIA, which can monitor the power supplied by the PCIe slot and 12V power supply.
In terms of graphics card temperature, the RTX 5000’s maximum temperature is 78.5°C, Omniverse 3D rendering is 75.9°C, and DaVinci Resolve 18’s 8K audio and video output is even lower with only 63°C; the software shows that the default GPU temperature limit is 84°C.
The graphics card’s TBP power consumption test achieved a maximum average power consumption of 250W in the Furmark stress test, while Omniverse 3D rendering maintained a power consumption of 237.5W and DaVinci Resolve’s 8K audio and video output 18 consumed only 131.9W.
↑ GPU temperature test.
↑ GPU power consumption test.
To summarise
The RTX 5000 professional graphics card uses the same AD102 core, 100 sets of SM memory and 32GB GDDR6 ECC, and GPU power consumption of 250W. It can also meet DaVinci Resolve 18’s 8K AV1 audio and video output and tracking real-time light UHD simulation rendering by Omniverse Creator. Generative AI’s stable diffusion text inference images all perform well.
The RTX 5000 is more affordable than the flagship RTX 6000 in terms of price and graphics card power consumption. It depends on the VRAM required for the project and the combination of multiple cards to choose a suitable professional graphics card; and professional graphics cards are 10.5 inches long and The 2-slot and fan workstation design requires only 250W of graphics card power consumption, making it easy to upgrade numerous workstations and servers.
NVIDIA provides professional users with a comprehensive software and hardware ecosystem, such as GPU virtualization, Mosaic, RTX desktop management, and integration of NVIDIA-developed SDK software tools, spanning artificial intelligence, data science, data centers, cloud, to design and simulation. robotics, high-performance computing, self-driving cars and other diverse fields, problems in all areas of life can be solved through GPU acceleration.
Finally, NVIDIA professional graphics cards come with comprehensive after-sales service from Leadtek Technology. Not only do they enjoy a 3-year home delivery warranty, but also have a toll-free customer service hotline “0800-600-206” to provide services in various fields. The most secure and stable creating and processing experience for users professional.
Welcome to leave your information here to learn more about NVIDIA. Exclusive Agent – Jieyuan will introduce you and serve you later. You can also be the first to receive the latest products and preferred information.
2024-01-03 08:17:10
#Unboxing #test #NVIDIA #RTX #Ada #Generation #professional #graphics #card #generation #audio #video #Omniverse #Creation #NVIDIA #RTX #Ada #professional #graphics #card #unboxing #tests #performance #LEADTEK #Omniverse