Design HPE solutions with a focus on AI and HPC workloads
A customer has developed a custom NIM and stores it in their NGC Private Registry. The customer wants to use that NIM in inferencing workloads running on HPE Private Cloud AI.
What should you recommend?
AFinding and importing NVIDIA AI blueprint that supports the custom NIM
BUsing MLIS with an NGC registry that points to the customer’s private registry
CUsing Ray Serve with custom-built deployments based on the custom NIM image
DDownloading the custom NIM to individual users’ models-pvc and manually creating KServe endpoints to deploy the NIM
0
Community Discussion
No comments yet. Be the first to start the discussion!
Design HPE solutions with a focus on AI and HPC workloads
An organization plans to run AI inference for customer service. It has strict environmental-efficiency standards and facility water access at the rack.
Which server configuration do you recommend?
AHPE ProLiant DL360 Gen12 with closed-loop liquid cooling
BHPE ProLiant DL580 Gen 12 with liquid-to-air rear-door cooling
CHPE ProLiant DL380 Gen12 with high-performance heat sinks
DHPE ProLiant DL380a Gen12 with 70% DLC
0
Community Discussion
No comments yet. Be the first to start the discussion!
Explain HPE compute AI and HPC solution components and architectureDesign HPE solutions with a focus on AI and HPC workloadsManage Customer Intent Documentation (CID) process to ensure the installation of the solution achieves the customer requirementsDemonstrate AISolutions
Explain HPE compute AI and HPC solution components and architecture
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
Save question
0
Community Discussion
No comments yet. Be the first to start the discussion!
What is one advantage of the NVIDIA AI Enterprise license included with HPE Private Cloud AI?
ADrivers to permit executing training pipelines on GPUs outside of containers
BAutoscaling of models across one or more GPUs based on current demand
CEntitlement to use NVIDIA components in blueprints for accelerating AI application development
DNVIDIA Data Center GPU Manager (DCGM) available from HPE Private Cloud AI platform
A customer runs an HPE Private Cloud AI solution with HPE AI Essentials 1.11. The customer has identified a NIM for an LLM they want to use and has imported NeMo microservices onto the AI Essentials platform. The customer intends to use NeMo Customizer’s LoRA to fine-tune the LLM for its needs.
What should you recommend the customer do to deploy the fine-tuned model on HPE Private Cloud AI using HPE Machine Learning Inference Software?
ASet up a custom OpenLLM registry in which they add the fine-tuned weights.
BStore the fine-tuned weights in the local S3 object store.
CUse environment variables when they deploy the model to point to the NeMo entity store running on AI Essentials.
DSubmit a support request to add the fine-tuned model to the NIMs hosted on HPE Private Cloud AI.
What benefit does the storage component provide in the small, medium, and large configurations of HPE Private Cloud AI?
AIt is built from the all-flash local drives on the AI-optimized nodes to provide the lowest latency.
BIt is built on GreenLake for File Storage, which supports both NFS and Lustre for mixed AI workloads.
CIt is built on GreenLake for File Storage, which supports low latency communications.
DIt is built from hybrid local drives on the AI-optimized nodes to balance performance and cost-effectiveness.
You are discussing HPE Private Cloud AI with a customer. The customer says they need a simpler way to visualize analyses of data spanning multiple database types, such as SQL Server and MongoDB.
What should you explain?
AThe customer can add these databases as data sources in HPE AI Essentials and then use that platform’s built in Superset and EzPresto frameworks to visualize the data.
BThe customer can add these databases as data sources in HPE AI Essentials. Users can then reference those sources in Knowledge Bases to run AI-based analytics and visualizations on them.
CThe customer can integrate databases as data volumes on HPE AI Essentials. The customer can then use Notebook servers included with HPE AI Essentials to browse and query the data.
DThe customer can synchronize these databases with GreenLake for File Storage, which is included as part of the Private Cloud AI solution. The customer can then use AI built into that solution to enrich the data with metadata and analyze it.
You propose 40 HPE Cray XD670 systems with NVIDIA H200 GPUs to a customer, along with HPE Performance Cluster Manager (HPCM).
What is one benefit of HPCM that you should emphasize?
AHPCM allows customers to monitor and optimize power consumption, including for NVIDIA GPUs.
BHPCM integrates with HPE Private Cloud AI to provide a single interface for monitoring system’s health.
CHPCM enables the systems to support hardware root of trust and OpenSSL.
DHPCM provides GPU-compatible scheduling software for HPC jobs.
A customer wants to use the Spark framework included with HPE Private Cloud AI to ingest and process data stored in an external MinIO object store. The customer wants to avoid distributing the store's access and secret keys to AI Users.
What should you recommend that AI administrators do?
AIntegrate the MinIO object store with the SSO and authorization framework provided by HPE Private Cloud AI’s Keycloak.
BCreate a custom spark image that is pre-integrated with the MinIO object store.
CSet up an HPE AI Essentials data source to the MinIO object store.
DReplicate the data from the MinIO object store to a NAS store compatible with the HPE AI Essentials Data Volumes format.
You are demonstrating HPE Private Cloud AI. A customer asks how the system manages multiple users who run concurrent workloads with different priorities.
Which feature would you emphasize to explain how this is handled?
AThe HPE AI Essentials scheduler plug-in
BThe solution’s Kubeflow pipeline architecture
CThe MLflow dashboard built into HPE AI Essentials
DThe built-in StorageClass and CSI driver
What is one advantage of the HPE-curated images for Spark supplied in HPE AI Essentials?
AThese images have been streamlined to reduce their size.
BThese images are ready to integrate seamlessly with HPE AI Essentials datasources.
CThese images were developed in conjunction with NVIDIA to interoperate with NIMs.
DThese images are pre-integrated with Jupyter tools to enable magic Spark commands.
A customer intends to use Ray with HPE Private Cloud AI and wants additional information about the framework’s integration with the solution’s GPUs.
What should you explain?
AThe solution’s Kuberay cluster has a GPU worker, and AI administrators can define the number of GPUs for that worker.
BThe solution’s Kuberay cluster supports GPUs by default and can autoscale up to the maximum number of GPUs on a single AI-optimized node.
CThe solution’s Kuberay cluster supports GPUs by default and can autoscale up to the maximum number of GPUs in the solution.
DThe solution’s default Kuberay cluster does not support GPU workers, but the customer can deploy an additional Kuberay cluster that does.
Which statement accurately describes the hardware security capabilities of HPE Cray XD systems?
AA BMC ASIC enables secure boot any root of trust capabilities similar to those in HPE iLO.
BThe HPE Cray XD systems include a sidecar node dedicated to security functions, such as monitoring firmware and OS for compromise.
CThe HPE Cray XD systems use the same iLO ASIC and firmware as the HPE ProLiant DL line.
DBMC software integrates with an iLO ASIC to provide hardware root of trust capabilities.
Which statement accurately explains how HPE servers are selected to serve as nodes in a VMware vSAN ESA cluster?
AYou can select vSAN ReadyNode configurations from the Broadcom Compatibility Guide or build your own configurations based on VMware guidelines.
BYou can build your own configuration based on VMware guidelines or select predefined HPE Smart Templates for vSAN.
CYou must select vSAN-compatible storage controllers from the Broadcom Compatibility Guide, but you make any choice for other hardware components.
DYou must choose an appropriate HPE vSAN ReadyNode based on the Broadcom Compatibility Guide.
You are preparing a proposal for HPE ProLiant servers to serve as vSAN nodes in a new ESA cluster.
What is one item of information required to determine the nodes’ drive-capacity requirements?
AThe RAID modes supported by the nodes’ storage controllers
BThe nodes’ iSCSI capabilities
CThe number of node failures to tolerate
DThe desired ratio of cache versus capacity drives
You need to demonstrate HPE Private Cloud AI running a complex AI application, such as a video search app, for a customer.
What is one useful resource for assembling this demonstration more quickly?
AHPE GitHub repo for Airflow demos
BThe tutorials folder in the HPE AI Essentials shared directory
CNVIDIA CUDA catalog
DNVIDIA AI blueprints
Which statement accurately describes the services for HPE ProLiant servers that use direct liquid cooling (DLC)?
AHPE Direct Liquid Cool Startup services are required for all racks of servers using DLC.
BHPE Direct Liquid Cool Startup services must be included for the first rack of servers that a customer deploys in a data center, but not for additional racks.
CHPE provides DLC consulting services and encourages partners to offer their own installation services.
DHPE provides optional DLC installation services, which partners can supplement with their own.
What is one reason to recommend deploying VMware vSAN OSA instead of ESA?
AThe customer wants to use features such as compression, which are not supported with ESA.
BThe customer wants to create a vSAN cluster that provides storage for other clusters but does not host VMs.
CThe customer wants a cost-effective solution that incorporates HDDs.
DThe customer prioritizes obtaining the highest write performance.
You are demonstrating HPE Private Cloud AI and HPE Machine Learning Inference Software to a customer. The customer says the organization has a custom model stored in an OpenLLM hub that it would like to deploy using HPE Machine Learning Inference Software.
What should you explain?
AHPE Machine Learning Inference Software supports pulling this model from the hub and deploying it.
BWhile HPE Machine Learning Inference Software does not support deploying custom models, Ray Serve, also provided by HPE Private Cloud AI, does.
CWhile HPE Machine Learning Inference Software does not support deploying custom models, KServe, also provided by HPE Private Cloud AI, does.
DHPE Machine Learning Inference Software can deploy this model if the customer packages it with Bento Archive and saves it to a PVC in HPE Private Cloud AI.
Company TXXX is developing autonomous-vehicle simulations and requires a system optimized for high-throughput training of generative AI models. The customer already uses HPE iLO to manage other servers and wants to retain that option.
Which HPE server model best aligns with these objectives?
AHPE Cray XD670
BHPE ProLiant DL580 Gen12
CHPE ProLiant Compute XD685
DHPE Cray XD2000
A customer needs to upgrade their ESXi hosts, and you are proposing HPE ProLiant DL servers. When you recommend using HPE custom ESXi images with these servers, the customer states that they include their own tools on all images.
What should you explain?
AThe customer can rely on the tools and drivers included in the HPE custom image to replace the functions of the tools the customer is currently using.
BThe customer can create their own custom image with their tools and with the HPE bundles available in the vibsdepot.
CThe customer should use an HPE OneView for vCenter OS Build Plan to combine their custom ESXi image with the HPE custom ESXi image.
DThe customer must use the custom image for ESXi to load successfully on the servers. However, the customer can add their tools to the OS manually after the server boots.
How does NVIDIA Spectrum-X handle congestion?
AIt uses flow metering based on information gathered from control packets to target congestion at the source.
BIt prevents congestion proactively by using ECMP to distribute traffic across multiple links.
CIt exclusively uses open standard technologies such as PFC and IP ECN to ensure compatibility for heterogeneous networks.
DIt allocates a dedicated buffer to each port to prevent traffic loss during congestion.
Community Discussion