Meteor Lake works economically | Diit.cz

If we don’t count the processor LakefieldIntel’s first attempt at a mobile chip made by layering different pieces of silicon using Foveros technology, is Meteor Lake Intel’s first tile product. Expectations were high as, in addition to deploying tiles, Intel promised a new large-core architecture, a new small-core architecture, and new manufacturing processes including Intel 4 and TSMC 3nm.

In the end, the new architectures turned out to be a refresh, which did not move the IPC in any way (in some situations there was even a reduction), the Intel 4 process achieves significantly lower clocks than the previous Intel 7, and the 3nm graphics were finally replaced by the slower 5nm, which although achieves competitive results in synthetic tests, but not in gaming rooms. The energy efficiency under load is also weaker compared to the competition, and the reduced consumption at rest (the effect of two small, very low-clocked cores, integrated into the central tile) managed to capture only a minimum of reviews. According to some sources, it is necessary to turn off applications running in the background and perhaps also some processes of the operating system, otherwise the processor tile will “wake up” and the energy advantage of two energy-saving cores will not manifest itself. According to some measurements, these cores are not even as efficient as expected, so the problem may lie elsewhere.

Even in terms of price, the situation is not very interesting – decently equipped laptops are significantly more expensive than competing or comparably equipped models built on the basis of an older generation (Raptor Lake) and those that are comparable in price to the competition are usually extremely stripped down (significantly worse displays and sometimes even a full-fledged USB 3.0 10 Gb/s is missing).

Although the whole situation could appear catastrophic for Intel, the company’s financial results for the first quarter show that the problem does not actually exist. It’s true that overall revenue didn’t do well and ended up below the pessimistic range of what Intel was officially expecting, but the laptops aren’t to blame:

Intel Q1 2024 Financial Results, Client Computing Group Left (Intel)

Although some media have pointed out how Intel’s client segment revenue has fallen, looking at the graph (see slide above, far left: CCG) it is clear that there was a drop only compared to the fourth quarter, which is the usual development (decline after pre-Christmas sales). On the contrary, the year-on-year development (Q1 2023 → Q1 2024) shows a very significant increase in income (by 29%) and an even greater strengthening of profits (by 117%). Given that Intel has released virtually no other new product for this segment to account for Q1’s performance, these results must be a significant dent Meteor Lake.

However, it is true that this may be the impact of the overall increase in demand that occurred after the post-Covid downturn subsided. People just buy more, upgrade, and for the most part, it doesn’t really matter what the processor in the new laptop is called.

AMD Q1 2024 Financial Results, Revenue Client Segment Left (AMD)

One cannot fail to notice that AMD also recorded a large increase in revenue (+85%) in the client segment, according to which the majority is due to interest in Ryzen 8000, in other words Hawk Point (Phoenix-refresh).

In addition to a significantly faster AI processor (10 TOPS → 16 TOPS), this one also brought an increase in energy efficiency, as shown by comparative tests compared to Phoenix (Ryzen 7000) in mobile consoles, where it made its debut. According to AMD’s CEO, however, even higher demand than after Hawk Point noted AMD after APU Strix Point (Zen 5 + RDNA 3.5), the release of which is confirmed for this year (unofficial sources estimate the second half of summer).