• Cisco UCSX-GPU-L4 L4 GPU gold heatsink with black bracket and fins
1 1

HPE

HPE S0K89C | NVIDIA L4 24GB GPU Accelerator, 72W, PCIe Gen4 x16

S0K89C

Stock Status: Enquire

Request Quote
Sale Sold out
Taxes included. Shipping calculated at checkout.

Description

The NVIDIA L4 24GB PCIe Accelerator delivers energy-efficient AI inference, video transcoding, and graphics acceleration in a compact low-profile form factor. Powered by Ada Lovelace architecture with fourth-generation Tensor Cores, this single-slot GPU consumes just 72W while providing 24GB GDDR6 memory and 300GB/s bandwidth. Ideal for edge deployments, virtualization, and high-density data center environments requiring powerful compute without specialized cooling infrastructure.

Features

Ada Lovelace GPU architecture with 4nm process technology
- Fourth-generation Tensor Cores with FP8 precision for 2x AI inference performance
- Third-generation RT Cores for hardware-accelerated ray tracing
- 24GB GDDR6 memory for large model inference and multi-tenant deployments
- Hardware AV1 encode/decode for efficient video streaming and transcoding
- NVIDIA DLSS 3 for AI-powered graphics upscaling and frame generation
- Low-profile, single-slot form factor for high-density server deployments
- Passive cooling design with 72W TDP for energy-efficient operation
- PCIe Gen4 x16 interface with 64GB/s bidirectional bandwidth
- Multi-instance GPU (MIG) support for virtualization and workload isolation
- Support for NVIDIA AI Enterprise software suite and containerized deployments
- CV-CUDA library acceleration for computer vision and video analytics
- CUDA, TensorRT, and Triton Inference Server compatibility
- Secure Boot and firmware update support via HPE iLO management
- NVIDIA vGPU software support for virtual desktop and application streaming

Warranty

All products sold by XS Network Tech include a 12-month warranty on both new and used items. Our in-house technical team thoroughly tests used hardware prior to sale to ensure enterprise-grade reliability.

All technical data should be verified on the manufacturer data sheets.

View full details

specs-tabs

Technical Specifications

FAQs

Technical Specifications

GPU: NVIDIA L4 Tensor Core GPU, Ada Lovelace architecture
- Memory: 24GB GDDR6, 192-bit memory interface, 300GB/s bandwidth
- Performance: 30.3 FP32 TFLOPS, 485 TFLOPS FP8 Tensor Core (sparse), 242 TOPS INT8 (dense)
- Tensor Cores: Fourth-generation Tensor Cores with FP8, FP16, BFLOAT16, TF32, INT8 support
- RT Cores: Third-generation RT Cores for ray tracing and rendering
- Video Engines: 2x NVENC, 4x NVDEC, 4x JPEG decoders with AV1 encode/decode support
- Form Factor: Low-profile, single-slot, half-height, half-length, passively cooled
- Interface: PCIe Gen4 x16 (64GB/s bidirectional bandwidth)
- Power: 72W maximum TDP, passive cooling
- Compatibility: HPE ProLiant DL320 Gen11, DL360 Gen11, DL380 Gen11, and other NVIDIA-Certified Systems
- Dimensions: Low-profile form factor (up to 9.5 inches length)
- Technologies: NVIDIA DLSS 3, CUDA, TensorRT, CV-CUDA, NVIDIA Omniverse support

FAQs

Q: What workloads is the NVIDIA L4 GPU optimized for?
A: The L4 is designed for AI inference at scale, video transcoding (especially AV1), virtual desktop infrastructure (VDI), cloud gaming, and lightweight graphics rendering. Its low 72W power envelope makes it ideal for high-density deployments where multiple GPUs are required per server without upgrading power or cooling infrastructure.

Q: What HPE ProLiant servers support the S0K89C NVIDIA L4 accelerator?
A: The S0K89C is compatible with HPE ProLiant DL320 Gen11, DL360 Gen11, and DL380 Gen11 servers. Installation requires a PCIe Gen4 x16 slot with support for low-profile adapters up to 9.5 inches length. High-performance fan kits may be required depending on server configuration and GPU count.

Q: How does the L4's 72W TDP benefit data center deployments?
A: The 72W thermal design power allows for higher GPU density per rack without requiring specialized power distribution or cooling upgrades. Organizations can deploy up to 8 L4 GPUs in compatible servers, significantly reducing total cost of ownership, rack space, and energy consumption compared to higher-wattage accelerators.

Q: Does the NVIDIA L4 require external power connectors?
A: No, the L4 draws all power through the PCIe Gen4 x16 slot and does not require auxiliary power connectors. The passive cooling design and low 72W power draw enable deployment in servers with standard power supplies and airflow configurations.

Q: What memory capacity and bandwidth does the L4 provide?
A: The L4 features 24GB of GDDR6 memory with a 192-bit interface delivering 300GB/s bandwidth. This capacity supports large model inference, multi-tenant workloads, and simultaneous video stream processing without memory bottlenecks, making it suitable for production AI and video applications.

Recently Viewed

  • Request a Quote

    Looking for competitive pricing? Submit a request, and our team will provide a tailored quote that fits your needs.

  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Call us Now  

  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Contact us