Intel® Xeon® 6 Processors vs. AMD EPYC for VDI: What Real World Testing Shows

Author: 

Richard Chuang, Principal Engineer, Intel

Co-Authors:

Bran Gavino, Segment Manager, Intel

Nehal Mehta, Director, Intel

Virtual Desktop Infrastructure (VDI) decisions are often influenced by headline specifications, especially core counts. Recently, this caused some to suggest that AMD-based platforms are better suited for VDI environments. But VDI success is not determined simply by core density or peak clock frequency. What matters most to IT teams and end users is consistent user experience, sustained performance under load, and cost per desktop at scale.

To understand how today’s platforms perform in real deployments, Intel® and HPE conducted industry-standard VDI testing using Login Enterprise, a widely adopted benchmark that simulates real user behavior and measures both scalability and experience quality.

The results are clear: Intel Xeon 6 processors on HPE ProLiant Gen12 servers deliver higher usable VDI density and better real-world efficiency than comparable AMD EPYC–based systems.(1)

Why VDI Is Different from Traditional CPU Benchmarks

VDI workloads behave differently from general compute benchmarks:

• End users care about responsiveness, not only peak throughput.
• Performance must be sustained, not bursty.
• IT teams size for headroom, not just maximum sessions
• Cost is driven by desktops per server, not cores per socket

This is why Login Enterprise is commonly used for VDI validation. It reports:

• VSImax: the maximum number of desktops supported.
• End User Experience (EUX): a combined score of responsiveness and resource utilization with a given number of active desktop sessions.

Apples to Apples Testing on HPE Infrastructure

The test environment compared similarly configured two-socket HPE servers running a Knowledge Worker workload:

• HPE ProLiant DL380 Gen12 with Intel Xeon 6 Processors (6730P, 6732P, 6517P and 6745P)
• HPE ProLiant DL385 Gen11 with AMD EPYC 9335
• HPE ProLiant DL380 Gen11 with 5th Gen Intel® Xeon® Processors (8562Y+) for baseline reference

All systems used the same VDI software stack, networking, and storage configuration to ensure fair comparison.

The Key Finding: Intel Xeon 6 Processors Support More Users per Server

Despite AMD’s higher advertised core performance, Intel Xeon 6 processors achieved up to 10%(2) higher virtual desktop density per server at comparable experience levels.

Login Enterprise Results

In practical terms, this means fewer Intel Xeon 6 CPUs are required to support the same number of users as AMD EPYC(5), which directly reduces capital, licensing, and operational costs.

Why Intel Xeon 6 Processors Perform Better for VDI

1. Sustained Clock Speed Matters More Than Core Count

VDI workloads, especially for knowledge workers and graphics-assisted use cases, are sensitive to per-core performance and latency. Intel Xeon 6 processors sustain all-core turbo frequency over 4.3 GHz under load(3), improving application launch times, UI responsiveness, and multi-tasking consistency. This aligns with HPE guidance that 3.0+ GHz clocks are critical for optimal VDI user experience(4).

2. Higher Density Lowers Total Cost of Ownership

With more desktops supported per server, organizations can:

• Reduce the number of physical servers
• Lower software and hypervisor licensing costs per desktop
• Cut power, cooling, and rack space requirements
• Simplify ongoing infrastructure management

3. More Headroom for Growth

VDI environments are rarely sized to their absolute limit. IT teams plan for growth and performance headroom. Intel Xeon 6 processors deliver higher desktop counts at “Excellent” (EUX≥8.0) and “Very Good” (EUX≥7.5) experience levels5 while having the headroom to accommodate more desktop sessions or more intensive workloads before hitting VSImax, allowing organizations to absorb user growth without immediately adding infrastructure or plan for disaster recovery scenarios.

4. Lower Risk Refresh for Existing Environments

For customers already running VDI workloads, Intel Xeon 6 processors enable live VM migration from older Intel platforms, minimizing downtime, avoiding desktop rebuilds, and reducing migration risk compared to platform changes that require full re-architecture.

Built for Today's VDI - and Tomorrow's Workloads

Beyond VDI performance, Intel Xeon 6 processors on HPE ProLiant Gen12 provide:

• Hardware-based VM isolation with Intel® Trust Domain Extensions (Intel® TDX)
• Built in AI acceleration with Intel® Advanced Matrix Extensions (Intel® AMX), without requiring GPUs
• HPE Silicon Root of Trust for zero trust infrastructure security

This makes the platform not only a stronger VDI choice today, but also a future-ready foundation for secure hybrid work and AI-enhanced desktops.

The Verdict is In

More cores do not automatically deliver better VDI. What matters is:

• How many users can you support per server?
• How consistent is their experience?
• How much infrastructure do you need to operate?

Testing with industry-standard tools and workloads shows that Intel Xeon 6 processor–powered HPE ProLiant Gen12 servers deliver higher usable VDI density5 than current-generation AMD EPYC solutions, helping organizations lower costs, reduce complexity, and deliver a better end-user experience.

Footnotes:

1,2,4: Based on external testing by HPE and AMD

3: See https://ark.intel.com for specifications

5: See “Configuration” for workloads and configurations. Results may vary.

Configurations

1-node, 2x Intel(R) Xeon(R) Platinum 8562Y+, 32 cores, HT On, Turbo On, NUMA 2, Total Memory 2048GB (16x128GB DDR5 5600 MT/s [5600 MT/s]), BIOS U54 v2.70, microcode 0x80003C0h, 1x Ethernet Controller E810-CQDA2, 13x 2.9T NVMe, VMware ESXi 9.0.0.0.24755229, 9.0.0.0, LoginEnterprise v6.3.14, 8C4G Win11 24H2 + Office 2021, Horizon 8 v2503, score=VM/host. Test by Intel as of 02/12/26

1-node, 2x Intel(R) Xeon(R) 6 6730P, 32 cores, HT On, Turbo On, NUMA 2, Total Memory 4096GB (16x256GB DDR5 6400 MT/s [6400 MT/s]), BIOS C220M8.4.3.4.462.0307251645, microcode 0x1000380, 1x Ethernet Controller E810-CQDA2, 13x 2.9T NVMe, VMware ESXi 9.0.0.0.24755229, 9.0.0.0, LoginEnterprise v6.3.14, 8C4G Win11 24H2 + Office 2021, Horizon 8 v2503, score=VM/host. Test by Intel as of 02/19/26

1-node, 2x Intel(R) Xeon(R) 6 6732P, 32 cores, HT On, Turbo On, NUMA 2, Total Memory 4096GB (16x256GB DDR5 6400 MT/s [6400 MT/s]), BIOS C220M8.4.3.4.462.0307251645, microcode 0x1000380, 1x Ethernet Controller E810-CQDA2, 13x 2.9T NVMe, VMware ESXi 9.0.0.0.24755229, 9.0.0.0, LoginEnterprise v6.3.14, 8C4G Win11 24H2 + Office 2021, Horizon 8 v2503, score=VM/host. Test by Intel as of 02/19/26

1-node, 2x Intel(R) Xeon(R) 6 6745P, 32 cores, HT On, Turbo On, NUMA 2, Total Memory 4096GB (16x256GB DDR5 6400 MT/s [6400 MT/s]), BIOS C220M8.4.3.4.462.0307251645, microcode 0x1000380, 1x Ethernet Controller E810-CQDA2, 13x 2.9T NVMe, VMware ESXi 9.0.0.0.24755229, 9.0.0.0, LoginEnterprise v6.3.14, 8C4G Win11 24H2 + Office 2021, Horizon 8 v2503, score=VM/host. Test by Intel as of 02/19/26

1-node, 2x AMD EPYC 9335, 32 cores, HT On, Turbo On, Total Memory 2304 GB (24x96GB DDR5 6400 MT/s [6400 MT/s]), BIOS 2.60 Virtualization-Max, 1x Ethernet Controller 25Gb, FC SAN – 32TB, VMware ESXi 8.0.3, 8.0.3, LoginEnterprise v5.13.6, 8C4G Win11 24H2, Horizon 8.16.0, score=VM/host. Test by HPE as of 2026 https://www.hpe.com/psnow/doc/a00155529enw

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.