HPE Private Cloud AI Solutions Sample Questions:

1. A customer needs a solution for AI inferencing at the edge for a computer vision application. They have determined a Small configuration of HPE Private Cloud AI is sufficient. The solution also requires several value-added services, including an AI transformation workshop and a data design service.
How should an architect attach these services in One Config Advanced (OCA)?

A) The architect must open a support ticket with HPE to have the services added to the quote.
B) The services must be quoted separately from the hardware and software configuration.
C) The services are automatically included in the Smart Template and require no action.
D) The services can be selected and added to the configuration as optional line items within OCA.

2. A customer is building a large-scale AI training cluster using HPE Private Cloud AI.
They have two main requirements for the storage solution:
1. It must provide extremely high throughput and low latency to feed data to a large cluster of NVIDIA H100 GPUs.
2. It must be able to scale performance and capacity independently to accommodate future growth without costly overprovisioning.
The architect is proposing HPE GreenLake for File Storage.
Which architectural features of this solution directly address the customer's requirements? (Select all that apply.)
```
Customer Requirements:
1. High performance for AI training workloads.
2. Independent scaling of performance and capacity.
```

A) It includes HPE Data Fabric, providing a global namespace across the worker nodes' local drives.
B) It supports GPUDirect Storage (GDS) over a high-speed NVMe-oF (RoCE) network, enabling a direct, low-latency path between storage and GPUs.
C) It uses a disaggregated architecture, allowing controller nodes (performance) and storage shelves (capacity) to be scaled separately.
D) It is managed exclusively through a command-line interface (CLI) for maximum control.
E) It relies on traditional hard disk drives (HDDs) for cost-effective capacity, with a small flash tier for caching.

3. An architect is designing a distributed training solution using HPE Private Cloud AI. The solution involves multiple HPE ProLiant DL380a servers, each with four NVIDIA H100 GPUs.
Review the logical network topology:
```
[HPE ProLiant Server 1] \
- GPU 0 <--> GPU 1 (Internal)
- GPU 2 <--> GPU 3 (Internal)
\
[NVIDIA Spectrum Switch (RoCE)] <--> [HPE GreenLake for File Storage]
/
[HPE ProLiant Server 2] /
- GPU 0 <--> GPU 1 (Internal)
- GPU 2 <--> GPU 3 (Internal)
```
Which interconnect technology facilitates the highest-bandwidth communication path labeled "(Internal)" for scaling the training job across the GPUs within a single server?

A) GPUDirect Storage (GDS)
B) PCI Express (PCIe) Gen5
C) RDMA over Converged Ethernet (RoCE)
D) NVIDIA NVLink Bridge

4. A customer states they need an AI solution for a "summarization" use case that will be accessed by approximately 75 concurrent users. The underlying data is updated frequently, so RAG will be required.
When using the HPE Intelligent Configurator, which three inputs are mandatory to get an initial sizing recommendation?

A) RAG requirement, Power and cooling capacity, Desired software license term
B) Use case, Number of users, RAG requirement
C) Number of users, Preferred GPU model, Data center location
D) Use case, Model choice, Network switch preference

5. An AI development team is training a large language model that exceeds the memory capacity of a single GPU. To continue training, they need to distribute the workload across multiple GPUs within the same server.
Which combination of HPE server and NVIDIA technology is specifically designed to provide a high- bandwidth, direct communication path between GPUs, bypassing the PCIe bus for improved performance in multi-GPU training?

A) HPE ProLiant DL380a Gen11 server with GPUDirect Storage (GDS)
B) HPE ProLiant DL380a Gen12 server with an NVIDIA NVLink Bridge
C) HPE Alletra Storage MP with RoCE-capable network adapters
D) HPE ProLiant DL325 Gen11 server with NVIDIA L4 GPUs

Solutions: