Skip to product information
1 of 2

Aruba

Aruba S0K90C | NVIDIA L40 48GB GDDR6 PCIe Gen4 GPU, 300W, Passive

SKU:S0K90C

Request Quote
Sale Sold out
Shipping calculated at checkout.

Stock Status: Enquire

20-min response

Need a quote or bulk pricing? Answered within 20 minutes, Sydney business hours.

Description

The NVIDIA L40 is a passively cooled, dual-slot, full-height PCIe Gen4 data centre GPU built on the Ada Lovelace architecture. It is aimed at data centre and edge server deployments running NVIDIA Omniverse, virtual workstations, cloud gaming, batch rendering, deep learning training and AI inference.

The card carries 48GB of GDDR6 memory with ECC across a 384-bit bus, giving 864 GB/s of memory bandwidth. Compute performance is rated at 90 TFLOPS FP32 and 724 FP8 Tensor TFLOPS with sparsity. It connects over a PCIe Gen4 x16 interface, draws up to 300W of board power through one PCIe 16-pin auxiliary power connector, and provides four DisplayPort 1.4a outputs. SR-IOV is supported with 32 virtual functions. No licence is needed for basic operation, while NVIDIA vGPU software is licensed separately if virtualisation is used.

The design relies on server chassis airflow rather than an onboard fan, so it needs a PCIe Gen4 x16 slot with adequate power and cooling. Listed supported HPE servers include the ProLiant DL320 Gen11, DL380 Gen10 Plus, DL380 Gen11, DL380a Gen11, DL385 Gen11, DL560 Gen11, ML350 Gen11 and XL675d Gen10 Plus, plus Superdome 3200, Superdome Flex and Superdome Flex 280.

Enquire with XS Network Tech for new or used pricing and availability.

Features

- Hardware-accelerated ray tracing and advanced shading
- AI capabilities for deep learning training and inference
- Part of the NVIDIA OVX server platform for Omniverse, graphics and simulation
- Supports virtual workstations, cloud gaming and batch rendering
- SR-IOV with 32 virtual functions and NVIDIA vGPU support
- Secure Boot and Root of Trust support
- ECC memory enabled by default, configurable in software
- Passively cooled design for data centre server airflow

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

Collapsible content

Technical Specifications

FAQs

Technical Specifications

- Part number: S0K90C
- GPU architecture: Ada Lovelace (AD102-895/895A)
- Memory size: 48GB GDDR6 with ECC
- Memory bandwidth: 864 GB/s
- Memory bus width: 384 bits
- Memory clock: 9,001 MHz
- FP32 performance: 90 TFLOPS
- FP8 Tensor performance: 724 TFLOPS (with sparsity)
- GPU clocks: 735 MHz base, 2,490 MHz boost
- System interface: PCIe Gen4 x16
- Maximum board power: 300W (300W default, 100W minimum)
- Power connector: One PCIe 16-pin auxiliary power connector
- Display outputs: 4x DisplayPort 1.4a
- Form factor: Full-height, full-length, dual-slot (4.4 in H x 10.5 in L)
- Cooling: Passive
- Board weight: 1,051 grams (excluding bracket and extenders)
- SR-IOV: Supported, 32 virtual functions
- Multi-Instance GPU (MIG): Not supported
- NVLink: Not supported
- Driver support: Linux R525 or later, Windows R525 or later
- CUDA support: CUDA 12.0 or later
- vGPU software support: vGPU 15.1 or later
- Operating temperature: 0 to 50 C
- Storage temperature: -40 to 75 C
- Operating humidity: 5% to 85% relative humidity

FAQs

Q: Does the NVIDIA L40 need a licence?
A: No licence is needed for basic operation. NVIDIA vGPU software, such as Virtual Compute Server Edition, is licensed separately if you use virtualisation.

Q: What power does this GPU need?
A: It has a maximum board power of 300W and requires one PCIe 16-pin auxiliary power connector. The server must supply enough power and a PCIe Gen4 x16 slot.

Q: Which servers is it listed as supporting?
A: Listed supported HPE servers include the ProLiant DL320 Gen11, DL380 Gen10 Plus, DL380 Gen11, DL380a Gen11, DL385 Gen11, DL560 Gen11, ML350 Gen11 and XL675d Gen10 Plus, as well as Superdome 3200, Superdome Flex and Superdome Flex 280.

Q: Does it have its own fan?
A: No. The card is passively cooled and depends on server chassis airflow, so it suits data centre servers designed for passive GPU cooling.

Q: Can two L40 cards be linked with NVLink?
A: No. NVLink and Multi-Instance GPU are not supported on the L40. Each card operates independently, with SR-IOV and vGPU available for virtualisation.

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