Which components make up an accelerated compute cluster?
Why C is the answer
GPU nodes, fabric, storage and management
A cluster needs compute, networking, storage and management.
In detail
- A cluster needs compute, networking, storage and management.
- A cluster is more than GPUs.
Use case
An illustrative scenario, not a real incident.
A team buys GPU servers but forgets the fabric and management, and the cluster cannot run well.
See it in different ways
flowchart TD A["GPUs in one server"] -->|"NVLink and NVSwitch"| B["Scale up"] B --> C["Server"] C -->|"InfiniBand or Ethernet"| D["Scale out to a cluster"]
mindmap
root((AI cluster))
Compute
GPU nodes
NVLink
Fabric
InfiniBand
Ethernet
Storage
High throughput
Management
Out-of-band networkmindmap
root(("Accelerated cluster"))
n1["Compute"]
n2["GPU nodes"]
n3["Network"]
n4["Fabric"]
n5["Data"]
n6["Storage"]
n7["Control"]
n8["Management"]Key point: Clusters need four parts.
Why the other options are wrong
- A. Only a storage array and a backup tape drive
- That leaves out compute.
- B. Only a web server and a load balancer
- That leaves out compute.
- D. GPU nodes alone, with nothing connected to them
- GPUs alone cannot work together.