NVIDIA L40S
Elevated AI and Graphic Performance

The Ultimate Universal GPU

Discover unparalleled multi-workload efficiency with the NVIDIA L40S GPU, a fusion of AI computing, top-tier graphics, and advanced media acceleration. Engineered for powering next-gen data center tasks, the L40S excels in generative AI, large language model inference and training, as well as 3D graphics, rendering, and video processing.

AMAX L40S Solutions

AceleMax DGS-428G2

4U 8x PCIe 5.0 GPU

CPUsAMD EPYC™ 9004 Series Processors
GPUs8x L40S
Memory24x DIMM slots
Storage8 x NVMe/SATA/SAS

AceleMax DGS-428G3

4U DP 8x PCIe GPU Server

CPUsDP 4th Gen Intel® Xeon® Scalable Processors
GPUs8x L40S
Memory32x DIMM slots
Storage12x NVMe/SATA/SAS

NVIDIA L40S

Unlock superior multi-workload capabilities with AMAX's NVIDIA L40S solutions, a full-height, full-length (FHFL), dual-slot 10.5-inch PCI Express Gen4 graphics solution based on the NVIDIA Ada Lovelace architecture, featuring a passively cooled design with a maximum board power of 350W. Designed for diverse applications, the L40S excels in hardware-accelerated ray tracing, AI features, advanced shading, and robust simulation, making it ideal for data center and edge server deployments in deep learning inference, generative AI training, batch and real-time rendering, virtual workstations, and cloud gaming.

RT Core Performance

212

TFLOPS

Tensor Core Performance

1,466

TFLOPS

Single-Precision Performance

91.6

TFLOPS


Ada Lovelace Architecture

The NVIDIA Ada Lovelace Architecture-driven L40S GPU boosts computational power for multiple industries, enhancing healthcare diagnostics and genome research, aiding financial risk analysis and fraud detection, and improving predictive maintenance in manufacturing. It accelerates autonomous vehicle development in the automotive sector with CAD and boosts visual effects in entertainment. The L40S GPU is pivotal for high-performance AI applications, driving innovation and competitive edge in the digital era.

Fourth-Generation Tensor Cores

Equipped with Fourth-Generation Tensor Cores, the L40S GPU pro-vides hardware support for structural sparsity and an optimized TF32 format, delivering immediate performance improvements for AI and data science model training. Each core enhances AI-driven graphics capabilities, utilizing DLSS to upscale resolution while optimizing performance in selected applications.

Third-Generation RT Cores

The Third-Generation RT Cores enhance throughput and enable concurrent ray-tracing and shading, improving graphical performance and accelerating rendering processes for product design, architecture, engineering, and construction projects. Experience lifelike designs with hardware-accelerated motion blur and captivating real-time animations.

Transformer Engine

The Transformer Engine in the L40S GPU accelerates AI performance and optimizes memory utilization for both training and inference phases. Leveraging the power of Ada Lovelace fourth-generation Tensor Cores, it scans the layers of transformer architecture neural networks, automatically recasting between FP8 and FP16 precisions, facilitating faster AI performance and expediting training and inference processes.


AMAX as your L40S Solution Provider

AMAX’s Elite partnership with NVIDIA facilitates a seamless deployment of the L40S GPU into your data center ecosystem. Our solutions architects will revolutionize your operations with a blend of AI compute, graphics, and media acceleration capabilities, turning data center challenges into competitive advantages. We design future-ready solutions that scale with your operational needs, keeping your IT infrastructure ahead of the technological curve.

Speak to one of our solution experts.Start engineering your ideal data center today.

© 2023 AMAX Engineering. All rights reserved.