Author: Lynn Comp, Head of Xeon Market Readiness, Intel Data Center Group
Scaling AI-driven applications poses many unique challenges for developers. For organizations running multi-modal AI applications that include text, images, video, and simulation data, the challenges increase even further. Fragmented infrastructure that separates AI, visualization, and simulation across different systems results in higher costs, inefficient resource use, and greater complexity when moving projects from pilot to production.
Oracle Cloud Infrastructure (OCI) has introduced an exciting solution to meet these very challenges. OCI Compute with NVIDIA RTX PRO is a new class of bare-metal VMs that delivers the high-performance compute, massive memory, and cost efficiency most needed to power today’s AI workloads, including multi-modal AI applications. OCI Compute with NVIDIA RTX PRO is built on the NVIDIA Blackwell architecture and powered by Intel Xeon 6 CPUs.
Intel Xeon 6 CPUs: Making the Difference
By pairing the OCI Compute with NVIDIA RTX PRO GPUs and Intel Xeon 6 processors, the CPU becomes a critical performance enabler for workload success. The difference lies in how Intel Xeon 6 processors improve data feeding, orchestration, memory movement, and overall efficiency.
- No More Bottlenecks
Intel Xeon 6 processors eliminate CPU bottlenecks feeding the GPU. This keeps the NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs fully utilized, reducing idle GPU cycles. This means faster inference pipelines, smoother real-time rendering, and better GPU ROI. - Higher Memory Bandwidth
Intel Xeon 6 processors bring ~33%-37% higher memory bandwidth (over previous generations) with up to 12 memory channels and faster DDR5/MRDIMM support. This translates to faster dataset feeds into the GPU.(1) - Built-in Inference Acceleration with Intel® Advanced Matrix Extensions (Intel® AMX)
Intel AMX provides matrix acceleration for AI inference on the CPU. This accelerator offloads preprocessing from the GPU and enables heterogeneous execution (CPU+GPU working together), especially important in agentic AI workflows and pre/post processing.(2) - Higher Core Density
Intel Xeon 6 processors have 2x the core density and higher per-core performance, which is especially important when running multi-modal and multi-service workloads.(3)
OCI Compute with NVIDIA HGX B300 Powered by 5th Gen Intel® Xeon® Processors
For organizations interested in training and frontier AI at scale, OCI offers a bare metal VM with the NVIDIA HGX B300 powered by Blackwell Ultra GPUs, one of top-tier Accelerators, powered by 5th Gen Intel Xeon processors. These are designed to handle trillion-parameter model training, large-scale Inference, Agentic AI, and RAG pipelines.
In addition to handling data pre-processing, orchestration, and scheduling, 5th Gen Intel Xeon processors include high PCIe bandwidth for GPU feeding, critical for large-scale training clusters.
5th Gen Intel® Xeon® processors also offer hardware-based security and a confidential computing environment through Intel® Trust Domain Extensions (Intel® TDX). Intel TDX creates Trust Domains (TDs) that isolate and protect the entire VM from the hypervisor, host OS, and other privileged software. Intel® Trust Authority is a SaaS that provides independent unified attestation of TDX using 5th Gen Intel Xeon host CPU and NVIDIA CC-enabled GPUs, allowing relying parties to establish trust in the confidential computing environment before sharing sensitive data or workloads.
OCI Compute with NVIDIA RTX PRO Has Changed the Model
Because this architecture enables AI inference, rendering, simulation, and media processing on a single system, deployment at scale is simplified, and separate infrastructure silos are eliminated, which removes overprovisioning and underutilization.
Customers can:
- Consolidate workloads onto a single GPU platform
- Improve resource utilization with concurrent workload execution
- Reduce infrastructure complexity and cost
- Deliver predictable, real-time performance for latency-sensitive applications
Support for Modern AI Workloads
OCI RTX PRO powered by Intel Xeon 6 processors support a wide range of workloads, including:
- Generative and AI Inference: RAG pipelines, Agentic AI, LLM finetuning.
- Visual Computing and Rendering: Media, Gaming, and Virtual production, Engineering, and Architecture workflows.
- Simulation and Digital Twins: Robotics, Synthetic data generation, Scientific computing, Engineering simulations.
Summary
In many GPU systems, CPUs can often be the bottleneck, limiting GPU utilization and increasing cost per workload…but not with Intel Xeon 6 processors on OCI Compute with NVIDIA RTX PRO.
With Intel Xeon 6 processors:
- GPUs get fed faster
- Bottlenecks are eliminated
- Intel AMX offloads AI tasks
- True heterogeneous computing happens
All of this results in higher GPU utilization, faster workloads and better overall price-performance.
Intel is excited to be OCI’s CPU of choice for many of their most important VMs. You’re invited to learn more about how OCI and Intel Xeon processors can take your AI efforts from pilot to production.
Learn More:
Oracle Cloud Releases New OCI X12 Standard Acceleron Instance Powered by Intel® Xeon® 6 Processors
Footnotes:
(1) https://blogs.oracle.com/cloud-infrastructure/announcing-oci-x12-standard-acceleron-compute
(3) https://blogs.oracle.com/cloud-infrastructure/announcing-oci-x12-standard-acceleron-compute
Notices and Disclaimers
Performance varies by use, configuration, and other factors. Learn more on the Performance Index site.
Performance results are based on testing as of the dates shown in configurations and may not reflect all publicly available updates. See backup for configuration details. No product or component can be absolutely secure.
Your costs and results may vary.
Intel technologies may require enabled hardware, software, or service activation.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.
여기에 의견을 추가하려면 등록된 사용자이어야 합니다. 이미 등록되어 있다면 로그인하시기 바랍니다. 아직 등록하지 않은 경우 등록 후 로그인하시기 바랍니다.