Intel shoots against AMD and puts itself in trouble

Intel’s marketing department had a creative moment. She created a presentation designed to educate PC users about “core truths”: “There is a long history of selling half-truths to unsuspecting customers. This guide is intended to help viewers understand some of those half-truths recognize that the competition is spreading it in front of everyone’s eyes,” wrote Intelbut has since taken the controversial document offline again.

Advertisement

Point 1 of the presentation: “The Ryzen 5 7520U is based on the outdated Zen 2 architecture released in 2019! […] The number 7 is the only new thing about some AMD 7000 series processors.” In one slide, the manufacturer goes so far as to compare AMD to a snake oil salesman – meaning the product is useless.

Intel draws attention to a valid point: The Ryzen 7000U/H notebook processor series is the pinnacle of creative naming in order to sell as many old processors as possible in a new look. The third digit should refer to the CPU architecture used (Zen 2, 3[+] or 4) draw attention, but how many of the millions of notebook buyers already know this?


(Image: Intel)

Things look completely different with Intel’s own processors. The Core i5-1335U as a concrete example uses the latest architecture (Raptor Lake), which was only presented in 2023. The problem? For the Core i-13000 and i-1300, Intel mixes new Raptor Lake dies with chips from the previous Alder Lake generation. Buyers may therefore receive a product from 2021.

To ensure that one Core i5-1335U is not faster than another, Intel artificially slows down the Raptor Lake types. Their biggest improvement is a larger Level 2 cache (2 vs. 1.25 MB), which allows the CPU cores to store more data locally. This means they have to access the slower RAM less often, which increases performance. However, Intel has eliminated the difference to make things equal.




The internet is full of hot IT news and stale pr0n. In between there are always pearls that are too good for /dev/null.

Many models from the 13th Core i generation therefore only have a small clock upgrade. The same thing with the 14th generation desktop CPUs: The Core i-14000 use the same dies as the Core i-13000, higher clock frequencies only bring marginally more performance to most models. Only with the Core i7 does Intel have more CPU cores activated.

Not to be forgotten is the fiasco with 10-nanometer production, which is why Intel has launched six CPU series based on the Skylake architecture – from the Core i-6000 (Skylake) to the Core i-10000 (Comet Lake) . Meanwhile, notebooks at least partially received 10 nm chips with newer architecture, which were mixed with 14 nm products.

The fourth “core truth” seems downright blasphemous: “Not all cores offer the best overall performance.” This statement comes from a company that has made hybrid computing socially acceptable with different CPU cores on desktop PCs and notebooks. To stay with the example of the Core i5-1335U: Intel and PC manufacturers advertise it as a 10-core, but only two performance cores achieve the highest performance. The eight efficiency cores are slower. To adapt, Intel even deactivates the AVX-512 instructions, which only govern the performance cores.

At this point we do not want to disparage the hybrid approach. Companies should just reconsider their statements about the competition if they are sitting in a glass house. Meanwhile, reactions to the presentation caused a swift backlash. Intel has removed the embed from its website. However, Google still finds a deep link to the PDF.


(mma)

To home page

source site