Table of Contents
Intel has unveiled the new Core Ultra 200 series for desktops codenamed Arrow Lake, in this article, we have covered it in detail, so read it till the end.
Performance with Intel Core Ultra 200 Series
Intel claims that with the Ultra 200 series the Lion Cove has gotten a 9% IPC uplift over the previous generation Raptor Cove CPU core, and in single-threaded benchmarks, it’s up to 8% faster than the 14900K CPU and up to 4% faster than the AMD 9950X CPU, and in multi-core benchmarks it’s up to 15% faster than the prior generation and up to 13% faster than the 9950X CPU. The Core Ultra 200 series CPUs also have a 3 MB L2 cache vs. 2 MB L2 for Raptor Lake, and it uses AI for power management.
Moving on to the e-cores, the new Skymont architecture brings a huge 32% IPC uplift vs the Gracemont architecture, which was present on the Raptor Lake, and the cores are 2x faster for AI throughput with the Intel Core Ultra 200 series.
The arrow lake ditches the hyper-threading so there is one thread for all the cores.
The Arrow Lake CPUs in the Ultra 200 series get a Xe-LPG-based architecture GPU with 4 xe cores and xess support with dp4a ai accelerated instructions; it has 4 MB of l2 cache with 4 ray tracing units and has full support for the Arc software stack with full support for DirectX 12 Ultimate; it has support for display port 2.1 and 4x 4k 60 monitors; and the media engine has full support for AV1 encoding and decoding.
Gaming
Now, with the Intel Core Ultra 200 series, let’s check out the gaming performance of the CPU. Intel claims that the CPU has gotten a giant leap in gaming performance perf/watt, mostly the same fps as Raptor Lake refresh but using 165w less system power. In Assassins Creed Mirage, 14900k gets an average of 264fps and 285k gets an average of 261fps. It is lower than the 14900k yes, but using 80 watts less system power than the 14900k, and Intel claims that they had slashed an average of 73W of power draw compared to the 14900K.
The lower wattages mean lower temps on board Intel claims that they saw an average of 13 degrees cooler CPU package with 360mm liquid cooler and 1080p gaming, and when Intel compared the 285K to the 9950X, they saw that in gaming we see that Intel lost to 9950X by 13% in the Cyberpunk 2077, and when we average the results of Intel losing, we found the 285K was losing by 9% average when compared to 9950X, but on some games, Intel wins by 15% average.
When compared to the 7950X3D CPU, Intel’s 285K falls behind by an average of 17%, but in the content creation tasks, the Intel Core Ultra 200 series has a huge lead up in their sleeves by winning the game by an average of 19% compared to the prior generation 7950X3D.
With the Ultra 200 series, Intel also claims that 265K is on average 5% slower than 14900k, 15 degrees cooler in Geomean, and consumes up to 188w less power.
AI, Content Creation, Efficiency
On the content creation side, Intel with their Core Ultra 200 series claims that their CPU is faster than the top-end and Ryzen 9950X in content creation, and Intel loses only in Adobe Lightroom Classic and Photoshop; the rest of them are Intel’s margin of victory.
Now coming to NPU sides, the NPU is based on Intel NPU 3 architecture, and it has 13 tops of AI power. The total platform can provide 36 platform tops, and the competition, which is 9950X, doesn’t have an NPU for AI-based tasks, though if you connect an A770 from Intel, it can provide up to 260 AI teraflops, which is faster than the NPU.
In geek-bench AI, with Ultra 200 series, Intel claims that the Arrow Lake CPUs can provide 2x performance vs. the Raptor Lake CPU, GPU, and NPU, and in Ul Procyon AI, Intel says that the Arrow Lake dominates the Raptor Lake in CPU, GPU, and CPU scores.
Intel claims that AI optimization fuels performance, and Intel says up to 50% faster AI performance than the competing processors.
On the efficiency side, in the Ultra 200 series, Intel has used the TSMC N3B node vs. the 14900K’s use of the Intel 7 node, which wasn’t that efficient, and Intel says that Arrow Lake draws up to 58% lower package power in lightly threaded works vs. the Raptor Lake refresh CPU, and Intel says it can give the same performance at half the package power.
Platform, Overclocking
Starting, with the Ultra 200 series, Intel has changed the socket from LGA1700 to LGA1851, and the chipset has been upgraded to the Intel 800 series. The platform has up to 24x PCIe 4 lanes with 10x USB 3.2,10x USB 2 ports and 8x SATA ports.
The platform also brings integrated Thunderbolt 4 to the desktop for the first time, and the platform also gets Thunderbolt Share for PC-PC screen sharing, quick data sharing to a new PC, peripherals, and storage sharing, and an integrated 1GB LAN port with integrated killer wifi 6E and Bluetooth 5.3 LE. It also has support for Thunderbolt 5 and 2.5GB LAN ports, Killer Wifi 7 and Bluetooth 5.4, but it depends on the manufacturer because it’s discrete.
Less move on to the overclocking, granular core clock top turbo frequency steps in 16.6 MHz for p and e cores, dual base clocks run independent bclk for compute and soc tiles, XTU includes automatic OC enhancements, new memory controllers support new XMP and CU-DIMM memory, low temperature overclocking increasingly bypasses voltage limits as the chip gets colder.
The platform gets support for up to ddr5-6400 you can overclock this memory via XMP for higher speeds and a max of 48 GB per dimm with 192GB of max RAM support with support for ECC RAM and adds support for CUDIMM And CSODIMM Memory
SKU and Cpu’s Availablity And Pricing
The sales and preorders for these core ultra 200 chips start on October 24, and 285K starts at 589 USD, with 265K starting at 394 USD. 265KF At 379 USD and 245K starting at 309 USD and 245KF at 294 USD.
In the SKU list, we can see that Intel has reduced the maximum turbo power for the core ultra 5 to 159w from 180w, the turbo speeds for the CPUs have been increased from their prior generation, and an improved GPU.
Arrow Lake H And HX
Intel says these CPUs will launch in Q1 2025, but they have shown us a sneak peek of the performance. Intel says that the Arrow Lake HX will have 24c 8 Lion Cove and 16 Skymont E cores with the same IGPU as the 285K with the same NPU.
The Arrow Lake H, however, comes with an upgraded XE LPG+ GPU with 4x AI throughput and 2x ray tracing and double the cache size with 8 MB L2 cache. It comes with 8 XE cores with 8 ray tracing units, and this has support for XE Matrix engines, which are optimized for AI and efficiency vs. the DP4A method that was used in HX CPUs, and this has up to 77 tops of GPU power with Intel AI playground access it has up to 99 platform tops of AI.
Conclusion
We hope this article related to the Intel Core Ultra 200 series was helpful to you. Continue reading GizmoGeek Hub for more tech news, stories, and reviews that will keep you up to date on the newest trends and advancements in technology. We attempt to provide in-depth and unbiased reviews, detailed tutorials, and up-to-date news to help you improve your tech knowledge and expertise.
Keep reading GizmoGeek Hub for more such articles & reviews.
Follow us on YouTube, X (previously Twitter), Instagram, LinkedIn, and our WhatsApp Channel to stay up to date on the latest news, announcements, and behind-the-scenes stuff.
Request – As a new site, we rely largely on advertising revenue to fund our operations. Please consider not using any ad blockers when viewing our website. Your support in viewing our advertisements is critical for us to continue providing high-quality content and services. We appreciate your understanding and cooperation in helping us keep GizmoGeek Hub running. Thank you for your support!
