HP NVIDIA RTX PRO 4000 Blackwell 24GB 4DP Graphics

HP NVIDIA RTX PRO 4000 Blackwell — Professional Graphics for Workstations

For users who work with large 3D models, complex scenes, or demanding AI applications, the HP NVIDIA RTX PRO 4000 Blackwell graphics card delivers top performance and efficiency. With Blackwell 2.0 architecture and a GB203 graphics processor built on 5nm technology, it handles even the most demanding workloads. With 24 GB GDDR7 memory and 672 GB/s bandwidth, the card ensures smooth operation without lag. The compact single-slot design and low power consumption of just 140 W allow for easy installation even in smaller workstations.

Architecture and performance for demanding applications

The graphics card is designed based on Blackwell 2.0 architecture using the advanced TSMC 4N FinFET 5nm process. This design not only delivers high performance, but also excellent energy efficiency. With 8,960 CUDA cores, 280 Tensor cores and 70 RT cores, the card excels in real-time rendering, AI and ray tracing applications. 36.83 TFLOPS (FP32) computing power ensures fast and smooth data processing even under high demands.

Connectivity and display support

With four DisplayPort 2.1b outputs, the card supports connecting up to four 8K displays simultaneously, which is ideal for professional multi-monitor work environments. This solution allows for maximum detail when working with graphics, visualizations and video. Support for the latest standards such as DirectX 12 Ultimate, OpenGL 4.6, Vulkan 1.4 and CUDA 12.0 ensures compatibility with a wide range of work and development tools.

Efficient design and low power consumption

The compact single-slot design with dimensions 241 × 111 × 20 mm is ideal for installation in smaller workstations. Consumption of just 140 W minimizes power requirements, while the card requires only one 16-pin connector. This makes it suitable even for systems with limited power sources, without having to sacrifice performance.

Technical data

• Graphics processor: NVIDIA GB203 (Blackwell 2.0)
• Manufacturing process: TSMC 4N FinFET, 5 nm
• CUDA cores: 8,960
• Tensor cores: 280
• RT cores: 70
• Memory: 24 GB GDDR7, 192-bit bus, 672 GB/s bandwidth
• GPU frequency: base 1,230 MHz, boost up to 2,055 MHz
• Memory frequency: 1,750 MHz (effective 28 Gbps)
• Outputs: 4× DisplayPort 2.1b
• Interface: PCI-Express 5.0 x16
• Consumption: 140 W (recommended power supply 300 W)
• Power: 1× 16pin connector
• Dimensions: 241 × 111 × 20 mm
• Slot: single-slot