Artikel Intel Xeon Diamond Rapids Packs 256 Cores and 1.28GB of Cache pertama kali tampil pada todayinasian.com.
]]>The details emerged following Intel’s presentation at Hot Chips 2026, where the company provided a deeper look at the architecture behind its upcoming Xeon 7 platform. Rather than simply increasing the size of a single processor die, Intel is embracing a highly modular design that combines multiple compute tiles with base tiles using advanced packaging technology.
Each compute chiplet can contain up to 16 P-cores, meaning the full 16-chiplet configuration reaches 256 cores. The design uses Intel’s 18A-P manufacturing technology for the compute silicon and incorporates Foveros Direct 3D packaging to connect the different components. This approach allows Intel to build a processor that would be extremely difficult to produce as one enormous monolithic die.
Read More: MacBook Pro M6 May Arrive Soon With a Major Chip Upgrade
The huge cache allocation is arguably just as striking. Diamond Rapids can provide up to 1.28GB of shared last-level cache, equivalent to 320MB per compute base block in the largest configuration. Keeping such a large amount of frequently accessed data close to the CPU cores could help reduce the need to repeatedly access system memory, particularly in demanding data-center workloads.
Memory and connectivity have also received major upgrades. Diamond Rapids supports up to 16 memory channels, with bandwidth reaching as much as 1.6TB/s using next-generation memory technology. The platform is also expected to offer 128 lanes of high-speed I/O, alongside support for PCIe 6.0 and CXL 3.0.
Intel’s architecture is aimed heavily at modern data-center workloads, including high-performance computing, cloud services and increasingly demanding AI-related applications. With hundreds of cores working together and enormous cache capacity, the processor is designed to keep large volumes of data moving efficiently through the system.
The FPS Review notes that Diamond Rapids represents a substantial architectural shift for Intel’s Xeon family rather than a simple specification bump. The company is expected to launch the new Xeon 7 generation in 2027, putting its massive chiplet architecture on course to compete for a larger share of increasingly complex server workloads.
Artikel Intel Xeon Diamond Rapids Packs 256 Cores and 1.28GB of Cache pertama kali tampil pada todayinasian.com.
]]>Artikel Intel Advances High-NA EUV for Panther Lake Chip Production pertama kali tampil pada todayinasian.com.
]]>High-NA EUV is regarded as one of the most advanced chipmaking technologies ever developed. Built by Dutch equipment manufacturer ASML, the system uses an upgraded optical design with a higher numerical aperture, enabling semiconductor manufacturers to print much finer circuit patterns onto silicon wafers. Compared with conventional EUV machines, the new technology offers significantly greater precision, reducing the number of manufacturing steps required while improving production efficiency and chip performance.
According to Interesting Engineering, Intel has become the first major semiconductor company to publicly demonstrate High-NA EUV systems operating as part of its manufacturing workflow. The technology is expected to play a crucial role in producing Panther Lake processors, which are scheduled for future commercial release. By simplifying complex lithography processes, Intel aims to increase manufacturing yields while lowering production costs for advanced semiconductor devices.
The company also reported encouraging results from its early production testing, indicating that High-NA EUV has met performance expectations during pilot manufacturing. Engineers believe the improved imaging capabilities will allow future processors to contain even more transistors within the same chip area, leading to higher computing performance, greater energy efficiency, and stronger artificial intelligence processing capabilities.
Read More: MIT Robot Swims and Flies Like a Diving Seabird Naturally
Discussing the achievement, Intel executives emphasized the importance of moving advanced research into large-scale manufacturing. “We are bringing High-NA EUV into high-volume manufacturing,” the company stated, describing the transition as a significant milestone for the semiconductor industry. The announcement highlights Intel’s long-term commitment to investing in next-generation manufacturing technologies that can support increasingly complex processor designs.
Beyond Panther Lake, High-NA EUV is expected to influence future generations of processors across personal computers, data centers, artificial intelligence systems, and high-performance computing applications. As transistor dimensions continue shrinking, advanced lithography will become increasingly essential for sustaining improvements predicted by Moore’s Law.
Looking ahead, as reported by Interesting Engineering, Intel’s progress demonstrates that High-NA EUV is moving from experimental development toward practical industrial deployment. While further optimization remains necessary before full-scale commercialization, the successful integration of this groundbreaking lithography technology represents an important advance in global semiconductor manufacturing, positioning Intel to compete more aggressively in the race to produce the world’s most advanced computer chips.
Artikel Intel Advances High-NA EUV for Panther Lake Chip Production pertama kali tampil pada todayinasian.com.
]]>