Er waren al een tijdje positievere signalen die geen wishfull thinking leken te zijn maar het begint er nu daadwerkelijk op te lijken dat Intel Foundry inderdaad de weg terug heeft gevonden.
TPU: Intel 18A-P Node Delivers 9% Performance Increase and 18% Power Savings
"Intel's next-generation 18A node is ready, and the company has tested it, showcasing some impressive results. At the VLSI 2026 Symposium in Honolulu, Hawaii, Intel will present its new research on the capabilities of the upcoming 18A-P node. According to the paper,
the 18A-P node can deliver a 9% performance increase at the same power level or achieve 18% power savings at the same performance level compared to the standard 18A.
However, there is more to this than meets the eye. Typically, node generations show similar performance and power improvements across generations.
What would be expected in terms of power and performance improvements when transitioning from 18A to 14A is now already available with the 18A-P node, but without any density improvements.
This makes the 18A-P node a very attractive option for external customers who expect the transistor density of the 18A node found in "Panther Lake," but with significantly better characteristics.
For reference designs, Intel uses an Arm core sub-block to test frequency and power scaling. The new 18A-P node can yield much better results on paper, but one of the most interesting improvements is in manufacturing, specifically in something called skew corners. When a node is manufactured, no two transistors are identical due to the inherent physics of the manufacturing process, especially at today's scale. These variations are measured between fast and slow "corners," meaning faster and slower transistors. The skew refers to how wide the performance and power gap is between these transistors.
Intel has managed to improve the skew corners on the 18A-P node by 30% compared to the standard 18A, meaning that power and performance characteristics are now more predictable, especially for parametric yields. This means that chip functions are now more predictable, and Intel has to deal with far fewer variations with the new node.
ETC.."
:strip_icc():strip_exif()/u/496922/Balance%252060x60.jpg?f=community)
/u/400/defember100.png?f=community)
/u/354532/crop619653ec8c6ca_cropped.png?f=community)
/u/202608/TURKIYE.JPG?f=community)
:no_upscale():strip_icc():strip_exif()/f/image/oSMyTeFdYm06YXbJXD0Dl1CH.jpg?f=user_large)
/u/46804/crop5f989efcbb253.png?f=community)