Top-of-the-line Intel® Core™ Ultra desktop processors (Series 2) and AMD Ryzen 9000 Series desktop processors are top contenders, as these processors offer powerful performance in gaming, video editing, and AI-enabled workloads.
These NPUs are specifically designed to run AI workloads efficiently, which can boost overall performance. AMD Ryzen 9000 Series desktop processors include some AI-enabled technologies that boost performance speeds but do not include an NPU.
Intel® Core™ Ultra desktop processors (Series 2) are part of Intel’s second generation of AI PC processors. With a new generation comes new features and additional performance-boosting technologies!
Intel® Core™ Ultra 9 processor 285K feature up to 8 Next-Gen P-cores which are optimal for AI or single threaded applications. These P-cores run at up to 5.7 GHz Max Turbo Frequency to deliver a speedy and powerful performance.
Intel® Core™ Ultra 9 processor 285K feature up to 16 Next-Gen E-cores which are optimal for multitasking or multi-threaded workloads. These E-cores have a Max Turbo Frequency of up to 4.6 GHz making them efficient and ideal for background tasks.
Workloads are automatically optimized through the Intel® Thread Director which schedules tasks to run on the most optimal cores.
Other noteworthy features of Intel® Core™ Ultra desktop processors (Series 2) are:
Let’s compare the performance of an Intel® Core™ Ultra 9 processor 285K and an AMD Ryzen 9 9950X desktop processor to see which offers a better performance in gaming and creating categories.
Powerful Intel® Core™ Ultra 9 processor 285K outperformed AMD Ryzen 9 9950X processors in select games , including many popular titles. When tested, the Intel® Core™ Ultra 9 285K processor provided:
See Notices & Disclaimers for additional information.
Content Creators can expect a small boost in video editing and animation performance with an Intel® Core™ Ultra 9 processor. When tested, the Intel® Core™ Ultra 9 processor 285K outperformed the AMD Ryzen 9 9950X processor in both content creation workloads having performed:
+6% FASTER editing videos in Adobe Premiere Pro (Pugetbench)
+4% FASTER animating content in Adobe After Effects (Pugetbench)
When tested, the Intel® Core™ Ultra 9 processor 285K consistently performed intensive single-threaded tasks better than the AMD Ryzen 9 9950X processor. The Intel® Core™ Ultra 9 285K processor is well equipped to handle demanding tasks, having performed:
+22% better (tested by SPECrate2017, a benchmark that measures computer performance and throughput using compute intensive application subtests)
+6% better (tested by Cinebench R23 ST, a benchmark from MAXON Computer that measures the performance of desktop 3D modeling applications)
+10% better (tested by Cinebench R24 ST, a benchmark from MAXON Computer that measures the performance of desktop 3D modeling applications)
On par performance (tested by Geekbench 6.3 SC, a benchmark tool that measures a device's CPU and GPU performance)
Intel® Core™ Ultra 9 processor 285K are built with 24 threads , making multi-threaded workloads a breeze for these desktop processors. When tested against an AMD Ryzen 9 9950X processor, the Intel® Core™ Ultra 9 processor 285K performed:
+22% better (tested by SPECrate2017_int_base (n-copies) (est.))
+13% better (tested by 3D Mark CPU Profile Max Threads)
+10% better (tested by Cinebench R24 MT)
+7% better (tested by Geekbench 6.3 MT)
Intel® Core™ Ultra processors have been intentionally designed to accelerate AI workloads and improve user efficiency. When tested against an AMD Ryzen 9 9950X processor, the Intel® Core™ Ultra 9 processor 285K performed:
Specially designed for AI workloads, Intel® Core™ Ultra desktop processors (Series 2) will utilize the NPU to intelligently boost your performance. As the first AI-accelerated desktop processors, Intel® Core™ Ultra desktop processors (Series 2) are built with Next-Gen P-cores and E-cores that are suited to handle your single and multi-threaded tasks.
Take the Quiz
You must view all content in this training to access the quiz.
For Retail Sales Professional training only. Do not repurpose or reuse without permission.
Performance results are based on testing as of dates shown in configurations and may not reflect all publicly available security updates. See backup for configuration details.
System Configurations
ARL-S & RPL-S 125W Configurations
Processors:
Intel® Core™ i9-14900K: RPL-SR 8+16+1
Intel® Core™ Ultra 9 processor 285K: ARL-S 8+16
Screen Size: 1920x1080
Video Card: Zotac Nvidia RTX 4090 – Driver: 32.0.15.6094
Motherboard: Asus ROG Maximus Z890 Hero: BIOS: 8001
Motherboard: Asus ROG Maximus Z790 Hero: BIOS: 2503
Memory ARL-S: G.Skill 2x16GB DDR5-6400 - 32-39-39-102
Memory RPL-SR: G.Skill 2x16GB DDR5-5600 - 28-34-34-89
Storage: Samsung 980 Pro PCIe NVMe SSD 1TB
Power Supply: EVGA 1200P2 Supernova
OS: Windows 11 Build 24H2 26100.1591
Water Cooling: Yes
Power Scheme: High Power
VBS: ON and Defender: ON
Granite Ridge 9000X Series 170W Configurations
Processor:
AMD Ryzen R9 9950X: 16C/32T, R9 7950X3D
Screen Size: 1920x1080
Video Card: Zotac Nvidia RTX 4090 – Driver: 32.0.15.6094
Motherboard: Asus Rog Crosshair X670E Hero: BIOS: 2401
Memory: G.Skill 2x16GB DDR5-5600 (9950X), 5200 (7950X3D) - 28-34-34-89
Storage: Samsung 980 Pro PCIe NVMe SSD 1TB
Power Supply: EVGA 1200P2 Supernova
OS: Windows 11 Build 24H2 26100.1591
Water Cooling: Yes
Power Scheme: High Power - VBS: ON and Defender: ON
Benchmark Details
CINEBENCH 2024 CPU is a benchmark from MAXON Computer that measures the performance of desktop 3D modeling applications. Reported metrics: CINEBENCH points (higher is better). Scaling efficiencies: CPU dominant, sensitive to FPU performance, frequency, core count and memory. OS support: Desktop Windows.
Puget bench for Lightroom Classic bench – this is a photo editing performance measurement benchmark developed by Puget Systems and is a part of Content creation benchmark suite. The benchmark can be accessed from: https://www.pugetsystems.com/labs/articles/PugetBench-for-Lightroom-Classic-1571/
Puget bench for Premier Pro – this is a video editing performance measurement benchmark developed by Puget Systems and is a part of Content creation benchmark suite. The benchmark can be accessed from: https://www.pugetsystems.com/labs/articles/PugetBench-for-Premiere-Pro-1519/
SPEC CPU2017 is a benchmark from the SPEC consortium (www.spec.org) that measures computer performance and throughput using compute intensive application subtests. SPECint_base2017 provides a comparison point as to how fast a device completes a series of single integer compute tasks. SPECint_rate_base2017 provides a comparison point for throughput, or how many integer compute tasks a device can accomplish in a given amount of time. SPECintrate_base2017 performance projections are based on Intel Compiler 2023.2.3 version.
Workload Details
Adobe Premiere Pro AI Audio Auto Tag workload: The AI Audio Auto Tag feature in Adobe Premiere Pro uses artificial intelligence to automatically categorize and tag audio clips based on their content. This feature can identify whether the audio is dialogue, music, sound effects, or ambient noise, and then assigns appropriate tags. This workload uses the Premiere Pro feature named 'Auto Tag' to analyze all the clips in a sequence and apply Audio Tags to them. The time to complete this workload is measured in seconds.
Audacity Music Generation workload: Audacity’s Music Generation allows users to create music directly within the software using AI. Powered by Intel's OpenVINO toolkit and Riffusion, this feature generates music based on user input, such as a text prompt describing the desired genre or style. This Workload uses the Audacity OpenVINO AI plugin to generate a 30 second music file using a prompt. The time to complete this workload is measured in seconds.
Topaz Photo AI 4X Upscale workload: This workload uses the 'Upscale and Resize' feature in Topaz Photo AI to apply a preset with 2X scale to one photo, then batch upscale the same settings onto (9) other files in the set. It measures the time in seconds to complete the 2X upscale on the (9) other images as and export all as JPEG at 100%. The source files are 1280 X 853 JPEG Files. The Preset is HighFidelityV1_10-90-100
Adobe Premiere Pro AI Auto Reframe workload: The Auto Reframe feature in Adobe Premiere Pro is an AI-powered tool designed to automatically reframe video footage to fit different aspect ratios. This workload uses the Premiere Pro feature named 'Auto Reframe' with the option for '1:1', to reframe an HD HEVC video source to 1:1. The time to complete this workload is measured in seconds.
Audacity Dynamic Noise Suppression workload: AI powered Dynamic Noise Suppression is a feature in Audacity introduced through Intel’s OpenVINO toolkit. It is designed to remove background noise from audio recordings, making it particularly useful for enhancing spoken word content like podcasts, voiceovers, or interviews. This workload uses Audacity’s Dynamic Noise Suppression feature to improve a 3-minute clip. The time to complete this workload is measured in seconds.
Productivity Performance while Collaborating workflow: This workload measures the time it takes to copy data from three workbooks to three worksheets, and all the pivot tables in the workbook. This task is completed while screen sharing in a 1x1 Zoom call (See “Zoom Collaboration workload” above).
Adobe After Effects Scene Edit Detection workload: The Scene Edit Detection feature in Adobe After Effects automatically detects and identifies scene changes within a video clip, making it easier to apply edits, effects, or transitions to specific scenes. This workload uses the After Effects feature named 'Scene Edit Detection' to analyze a video and apply cut points to each unique scene. The time it takes to complete this workload is measured in seconds.
Multitasking Creator workflow: This concurrent multitasking creator workflow leverages Adobe Premiere Pro in the foreground, along with Blender in the background. In the foreground via Premiere Pro, the creator is working on a new video project to later upload online. They playback their timeline which contains the XAVC-H video, to review and plan video edit requirements. While this happens, in the background they render an animation of their channel’s logo, to later be pulled into their Premiere Pro project. The foreground and background times are measured in seconds.
Adobe Lightroom Classic AI Raw Details workload: The AI-powered "Raw Details" feature in Adobe Lightroom Classic enhances the quality of raw image files. This workload uses the Lightroom Classic feature from the 'Enhance' tools named 'Raw Details' to apply an effect to (10) Raw images. (5) CR2, (4) NEF, (1) CR3. The time to complete this workload is measured in seconds.
Adobe Premiere Pro 8K HEVC Playback workload: This workload measures the playback framerate in FPS, of 8K HEVC 24fps 124mbs source video in Adobe Premiere Pro timeline at full resolution.
Adobe Premiere Pro 5.7K DNxHR Playback workload: This workload measures the playback framerate in FPS, of 5.7K DNxHR 60fps 751mbs source video in Adobe Premiere Pro timeline at full resolution.
Adobe Premiere Pro 8K XAVC-H Playback workload: This workload measures the playback framerate in FPS, of 8K XAVC-H 29.97fps 140mbs source video in Adobe Premiere Pro timeline at full resolution.
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.