Introducing NVIDIA RTX™ A6000 GPU Instances on Lambda Cloud
Blog post from Lambda
NVIDIA RTX A6000 GPU Instances are now available on Lambda Cloud, a first public cloud to offer instances with 1x, 2x, and 4x NVIDIA RTX A6000 GPUs. These instances provide great price-to-performance value, with ~2x average performance improvement over previous instances, 48 GiB of VRAM per GPU, NVLink support for faster GPU-to-GPU communication, and up to 200 Gbps of inter-node bandwidth. The new instances are powered by AMD EPYC CPUs and offer high-speed NVMe storage, making them suitable for general-purpose, high-memory, and multi-node workloads. Pricing starts at $1.00/hr for a single GPU instance, with discounts for larger configurations. The new instances provide significant performance improvements over on-demand p3.8xlarge instances and even existing NVIDIA RTX 6000 instances, offering an average of 150% and 10% better performance respectively.