What is TensorDock?
TensorDock is a cloud service provider offering a global fleet of GPU and CPU servers. It specializes in providing infrastructure for deep learning, AI, rendering, and scientific computing workloads. The platform emphasizes affordability, offering up to 80% cost savings compared to traditional cloud providers.
TensorDock features a wide variety of GPU models, including NVIDIA H100, A100, and RTX 4090, as well as the latest Intel Xeon and AMD EPYC CPUs. It employs KVM virtualization, providing users with root access and dedicated resources. The platform ensures 99.99% uptime, and robust security measures, including vetting hosts, restricting hostnode access and monitoring the nodes.
Features
- Industry's Widest GPU Selection: Offers 45 GPU models, from RTX 4090 to HGX H100 SXM5 80GB.
- KVM Virtualization: Provides root access, dedicated GPUs, and full OS control.
- Competitive Pricing: Up to 80% less expensive than other cloud providers.
- Global Scale: GPUs distributed across 100+ locations in over 20 countries.
- Reliable API: Well-documented API for managing servers.
- Bulletproof Reliability: 99.99% uptime standard and vetted hosts.
Use Cases
- Deep learning model training and inference
- AI development and deployment
- High-performance computing (HPC)
- Scientific computing and simulations
- Image processing and rendering
- Cloud gaming
- Transcoding and batch processing
FAQs
-
How do you guarantee the security of customer data?
Most hardware is hosted in certified data centers. SSH access is revoked from hosts, and hostnode access is restricted to authorized personnel only. An agent monitors hostnodes for logins and suspicious activity. -
How does billing work?
TensorDock operates on a pay-as-you-go model. Users deposit funds, and the balance is deducted continuously after a server is deployed. Servers are automatically deleted when the balance reaches $0. -
Why does the same GPU/CPU have multiple prices?
TensorDock is a marketplace of independent hosts who compete and set their own pricing. Additionally, hosts have different locations and redundancy, leading to varying costs.
