Universal usage
The L40S GPU Instance offers unparalleled performance across a spectrum of tasks, including gen AI, LLM inference, small-model training and fine-tuning to 3D graphics, rendering, and video applications.
L40S GPU Instance
Accelerate the next generation of AI-enabled applications with the universal L40S GPU Instance, faster than L4 and cheaper than H100 PCIe.
Cost-effective scalability
Starting at €1.4/hour for 1 GPU with 48GB of GPU memory and available in 4 different formats (1, 2, 4, 8 GPUs), the L40S GPU Instance enables cost-efficient scaling according to workload demands, ensuring optimal resource utilization on top of high-performance capability.
K8s compatibility
Seamlessly integrate the L40S GPU Instance into your existing infrastructure with Kubernetes support, streamlining deployment and management of AI workloads while maintaining scalability and flexibility.
Available zones:
Paris:PAR 2L40S GPU technical specifications
GPU NVIDIA L40S GPU
GPU memory48GB GDDR6 (864GB/s)
Processor8 vCPUs AMD EPYC 7413
Processor frequency2.65 Ghz
Memory92GB of RAM
Memory typeDDR4
Network Bandwidth2.5 Gbps
Storage1.6TB of Scratch Storage and additional Block Storage
CoresTensor Cores 4th generation RT Cores 3rd generation
Ideal use cases with the L40S GPU Instance
LLM fine-tuning & training
Use H100 PCIe GPU Instances for Medium to Large scale Foundational model training but harness L40S capabilities to fine tune in hours and train in days small LLMs.
- An infrastructure powered by L40S GPUs can train Models in days
To train Llama 2-7B (100B tokens) it would require 64 L40S GPUs and take 2.9 days (versus 1 day with H100 NVlink GPUs, like on Nabu2023) - Fine-tune Models in hours
To fine tune Llama 2-70B SFT (1T tokens), it will require 64 L40S GPUs L40S and take 8.2 hours (versus 2.5hours with H100 NVlink GPUs, like on Nabu2023)
Source: NVIDIA L40S Product Deck, October 2023
Speed up Llama 2-7B inference with 1 L40S GPU
L40S GPU Instance, with 1 GPU, delivers inference performance that enables you to put Llama 2-7B main use cases in production efficiently:
- Chat: generate immediate responses to prompts with context
A L40S* delivers a response to an input sequence of 225 with an output of 20 in 16msec for 1st token latency and 246 msec for End-to-end latency - Q&A: generate quick responses to short questions
L40S* delivers a response to an input sequence of 16 with an output of 64 in 12msec for 1st token latency and 771 msec for End-to-end latency - Summarization and RAG: retrieve augmented responses based on interpretation/summarization of multi page content
L40S* delivers a response to an input sequence of 2048 with an output of 128 in 77msec for 1st token latency and 1703 msec for End-to-end latency
*serving a LLama 2-7B model in FP8
Take Graphic experience to the next level
Rendering is essential for bringing visual elements together, from lighting effects to textures, to generate the final result. Leveraging the Lovelace architecture, the L40S GPU Instance demonstrates superior performance compared to the Ampere architecture, particularly in rendering and graphics tasks. Some concrete results for you to compare:
- Rendering: the L40S is up to 2.4 times faster than a A40 for rendering (on Blender Cycles render engine)
- Graphics: the L40S is up to 2.3 times faster than a A40 for graphics (Graphics: Viewperf2020; snx composite 1080p, energy composite 4k)
Source: NVIDIA L40S Product Deck, October 2023
Build and monitor a flexible and secured cloud infrastructure powered by GPU
Benefit from a complete cloud ecosystem
Kubernetes Kapsule
Match any growth of resource needs effortlessly with an easy-to-use managed Kubernetes compatible with a dedicated control plane for high-performance container management.
Load Balancer
Distribute workloads across multiple servers with Load Balancer to ensure continued availability and avoid servers being overloaded.
Virtual Private Cloud
Secure your cloud resources with ease on a resilient regional network.