DGX+Parallel Storage Building Blocks for Scale Out AI
DGX-POD is an NVIDIA-validated building block of AI Compute & Storage for scale-out deployments.
Designed for the largest datasets, DGX POD solutions enable training at vastly improved performance compared to single systems. DGX-POD also includes the AI data-plane/storage with the capacity for training datasets, expandability for growth, and the speed that can keep up with AI workloads.
Trust Our Partnership for Your DGX-POD Delivery: Microway + RAID Inc.
NVIDIA DGX
Experts in AI & HPC compute, Microway delivers the fully integrated DGX AI Appliances for complete DGX POD solution integration
Parallel Storage
Experts in AI & HPC storage, RAID Inc. delivers the parallel storage from DDN connected to the DGX-1 or DGX-2 appliances in our DGX PODs
Why DGX-POD?
Fully Integrated Deployment
DGX-POD solutions come with the DGX AI Appliance and Parallel storage fully integrated together
Extendable
Build complete racks with multiple DGX-POD configurations
Validated Configuration
DGX-POD solutions have been validated to deliver high storage throughput
Performance Improves Over Time
DGX-POD automatically receives performance improvements from NVIDIA
DGX-1 Solutions
- 1:1 DGX-1 POD with AI200 Storage
- 4:1 DGX-1-POD with AI200 Storage
- 9:1 DGX-1-POD with AI400 Storage
- Something else?
1 NVIDIA DGX-1 with DDN AI200 Delivered by RAID Inc.
First Deployment of AI Compute & AI Ready-Parallel Storage
Specifications
- 1 Tensor PFLOPS of AI Performance
- DDN AI200 appliance with throughput up to 25GB/sec and 750,000 IOPS (various data capacities available)
- Total of 256GB of GPU memory
- Mellanox 100Gb EDR InfiniBand
- NGC Containers with NVIDIA-optimized performance
- Full parallel filesystem and DDN Management GUI
- Validated for Linear Scaling: up to 6000 imgs/s on Resnet50, and linear performance scaling when adding containers
- Superior Performance nearly 2X the training data throughput of competing NFS-based appliances!
- Seamless AI throughput scaling: add DGX appliances!
- Seamless storage throughput & capacity scaling: add AI200 appliances and filesystem is ready to grow or deliver greater bandwidth
4 NVIDIA DGX-1s with DDN AI200 Delivered by RAID Inc.
Scale-Up AI Compute & AI Ready-Parallel Storage
Specifications
- 4 Tensor PFLOPS of AI Performance
- DDN AI200 appliance with throughput up to 25GB/sec and 750,000 IOPS (various data capacities available)
- Total of 1TB of GPU memory
- Mellanox 100Gb EDR InfiniBand
- NGC Containers with NVIDIA-optimized performance
- Full parallel filesystem and DDN Management GUI
- Seamless AI throughput scaling: add DGX appliances!
- Seamless storage throughput & capacity scaling: add AI200 appliances and filesystem is ready to grow or deliver greater bandwidth
9 NVIDIA DGX-1s with DDN AI400 Delivered by RAID Inc.
Full Rack AI Compute & AI Ready-Parallel Storage
Specifications
- 9 Tensor PFLOPS of AI Performance
- DDN AI400 appliance with throughput up to 50GB/sec and 1.5 million IOPS (various data capacities available)
- Total of 2.25TB of GPU memory
- Mellanox 100Gb EDR InfiniBand
- NGC Containers with NVIDIA-optimized performance
- Full parallel filesystem and DDN Management GUI
- Seamless AI throughput scaling: add DGX appliances!
- Seamless storage throughput & capacity scaling: add AI400 appliances and filesystem is ready to grow or deliver greater bandwidth
Custom Parallel Storage Solutions
Looking for another filesystem (Spectrum Scale/GPFS, BeeGFS), scale, or capacity? Let us design a system that meets your needs.
Key Capabilities
- Capacities to 1PB or beyond
- Throughput >100-500GB/sec
- Dynamic capacity expansion
- Spectrum Scale (formerly GPFS), BeeGFS, or Lustre
DGX-2 Solutions
- 1:1 DGX-2 POD with RAID Inc. AI200 Storage
- 3:1 DGX-2-POD with RAID Inc. AI400 Storage
- Something else?
DGX-2 with DDN AI200 Delivered by RAID Inc.
NVIDIA DGX-2 Appliance & AI-Ready Parallel Storage
Specifications
- 2 Tensor PFLOPS of AI Performance
- DDN AI200 appliance with throughput up to 25GB/sec and 750,000 IOPS (various data capacities available)
- Total of 512GB of GPU memory
- Mellanox 100Gb EDR InfiniBand
- NGC Containers with NVIDIA-optimized performance
- Full parallel filesystem and DDN Management GUI
- Seamless AI throughput scaling: add DGX appliances!
- Seamless storage throughput & capacity scaling: add AI200 appliances and filesystem is ready to grow or deliver greater bandwidth
3 NVIDIA DGX-2s with AI400 Delivered by RAID Inc.
Full Rack AI Compute & AI Ready-Parallel Storage
Specifications
- 6 Tensor PFLOPS of AI Performance
- DDN AI400 appliance with throughput up to 50GB/sec and 1.5 million IOPS (various data capacities available)
- Total of 512TB of GPU memory
- Mellanox 100Gb EDR InfiniBand
- NGC Containers with NVIDIA-optimized performance
- Full parallel filesystem and DDN Management GUI
- Seamless AI throughput scaling: add DGX appliances!
- Seamless storage throughput & capacity scaling: add AI400 appliances and filesystem is ready to grow or deliver greater bandwidth
Custom Parallel Storage Solutions
Looking for another filesystem (Spectrum Scale/GPFS, BeeGFS), scale, or capacity? Let us design a system that meets your needs.
Key Capabilities
- Capacities to 1PB or beyond
- Throughput >100-500GB/sec
- Dynamic capacity expansion
- Spectrum Scale (formerly GPFS), BeeGFS, or Lustre
