Intel’s latest Arrow Lake processor for desktop PCs combines the Meteor Lake and Luna Lake architectures, taking over Meteor Lake’s NPU and Luna Lake’s abandonment of hyperthreading. Hyper-threading is also missing this time for similar reasons as the hyper-threading feature was excluded from Luna Lake.
Arrow Lake, also known as the Core Ultra 200S processor lineup, is Intel’s first ‘breakdown’ desktop processor and is built based on tiles, meaning each part of the chip is manufactured individually in a different process. Intel released in-depth information about Luna Lake’s architecture and the model, price, and performance of the Core Ultra 200S processor. The key omission is hyperthreading, just like the Intel Luna Lake mobile processor.
The advantage of Arrow Lake is simple. It has better performance than the 14th generation Core chip while significantly reducing power consumption. To achieve this, Intel executives said they applied the same thinking to both Luna Lake and next-generation desktop processors: making cores as efficient as possible in both power and space.
What is hyperthreading?
Hyperthreading (also known as simultaneous multithreading) is a very simple concept. Each processor core is designed to execute one thread of instructions, but hyperthreading creates a second ‘virtual’ processor inside a single processor core. Hyperthreading allows individual processor cores to always be running, executing instructions on at least one of the two cores. The last thing enthusiasts want is for CPU cores to sit idle when they could be doing useful work.
The problem is that the second core is a virtual core and not a ‘true’ second physical core. This can lead to resource contention and additional overhead, and the question of whether or not to turn on hyperthreading while gaming has been a topic of debate for years.
Meanwhile, Intel has been going back and forth on this feature: Some of Intel’s 9th, 10th, and 11th generation Core processors exclude hyperthreading, such as the Core i7-9700K, while Intel’s Atom chips use hyperthreading. I didn’t. However, most Intel Core chips use hyperthreading. However, AMD has consistently used hyperthreading and is still using it. There was always a problem. Can hyperthreading deliver performance gains that exceed the system latency, controller die cost, and power consumed by hyperthreading?
With Luna Lake, the answer was no, and that continues for Intel’s latest desktop chips. In part, this is because Arrow Lake borrows heavily from Luna Lake, using the same Lion Cove performance core and Skymont efficiency core that appeared in Luna Lake.
Intel is ahead in terms of power and performance because it doesn’t use hyperthreading by default, said Robert Hollock, Intel’s vice president and general manager of client AI and technology marketing. Arrow Lake includes both desktop and mobile processors, and Hollock was fielding questions about implementing desktop hyperthreading. But Hollock’s answer appears to apply to both desktop and mobile chips.
“It’s really a combination of several things,” Hollock said. First, we found that we could actually save power consumption by not including hyper-threading in the product, and we can see that even without hyper-threading, we are ahead in multi-core performance by about 15% to 20%. Therefore, it is of the opinion that the goal was achieved in terms of overall computing performance while increasing efficiency.
Hollock added, ‘Another thing is that it is the same design used in Luna Lake.’ “We were able to take these cores and designs and immediately integrate them using Intel Foveros technology. “Two things influenced our decision: speed to market and maximizing performance per watt.”
editor@itworld.co.kr
Source: www.itworld.co.kr
