Deploy GPUs

Overview

---

Access GPU deployment solutions on-demand for AI, ML, HPC, and enterprise-grade applications. This section guides you through leveraging high-performance GPU infrastructure with Virtual Machines or Bare Metal, available whenever you need them.

• Deploy and manage GPU resources.

• Understand the differences between Virtual Machines and Bare Metal to choose the best fit for your workload.

• Scale your GPU infrastructure on-demand, whether you need flexibility or raw performance.

Node Specifications

---

H100 SXM5

GPU Configuration	8x NVIDIA H100 SXM5
CPU Configuration	2x Intel Xeon Platinum CPUs
Memory	1 TB DDR5 ECC RAM
Storage	19.2 TB NVMe PCIe (5x 3.2 TB data disks + 1x 3.2 TB boot disk)
Public IP bandwidth	100 Gbps (Best Effort)

Networking Options

---

Virtual Machine (VM)	Bare Metal
100 Gbps Ethernet	100 Gbps Ethernet
	3.2 Tbps InfiniBand

Our Virtual Machines (VM) instances are only available in 100 Gbps Ethernet.

GPU Deployment Solutions

---

Virtual Machines (VMs)

• Pre-booted and ready for near-instant deployment.

• Ideal for smaller projects or workloads requiring quick and flexible scaling.

Bare Metal

• Dedicated hardware with full control and on-demand provisioning.

• Perfect for large-scale workloads or environments needing advanced customization.

---

This ensures you can spin up or scale down resources as needed, with no long-term commitments or delays.