How strange was this presentation conference of the 11e generation of Core Mobile processors! While Intel generally avoids citing the competition too bluntly, preferring the (not very humble) generational comparisons between its own chips, the x86 processor champion has shown its teeth this time around.

The day before last night, at the official launch of its 11e generation of “Tiger Lake” mobile processors, the firm has therefore released performance measurements comparing its feline not only to AMD’s competing mobile SoC – the Ryzen 7 4800 U – but also with the Nvidia Geforce MX350 graphics chip, a GPU that found in entry-level machines and which represents the first price to play in 3D.

And took advantage of these demos to qualify the gaming experience offered by its competitors, depending on the title, as “barely playable”. A speech rarely seen in the very polished communication of the mastodon.

Video games, authoring software, audio filters accelerated by AI, Intel did not use the back of a spoon in its examples to highlight, but above all to compare its chip with its competitors. Because Intel has faith in its chip, of course. But also because Intel is a little annoyed …

Intel is tired of taking hits

If you’ve read our news about Tiger Lake, you already know the basics of this new SoC. A “super” processor for laptops which not only represents a nice evolution on the CPU side, but which also promises a doubling of graphics performance compared to Ice Lake – ie a fourfold increase in two generations! Besides the pride in launching such a major new component – Intel has been an outrageously dominant player in the laptop market forever – what explains Intel’s (gently) vindictive attitude is the context.

Entangled in production concerns which saw him accuse of delivery problems on many ranges and who delayed the introduction of its 7 nm engraving, the giant Intel, which was once the champion of fine engraving, has since ceded the crown to TSMC and Samsung. And he may explain that the measurement of transistors is not standardized between the different players, highlighting fairly unique technologies – EMIB, Foveros, SuperFin – nothing helps: for the whole world, Intel has been abandoned.

Lost in the desert for years (notably) because of its Bulldozer architecture, AMD has found its way north thanks to its extraordinary CEO, Ms. Lisa Su. And has a series of successes with its Zen (CPU) and RDNA (GPU) architectures. Not content with being in command of the next two generations of consoles (Xbox Series X and Playstation 5) with its chips, AMD dominates the price / performance ratio of desktop processors. And its latest generation of x86 laptop chips, the Ryzen 4000 Mobile, is a clear success which allows it to grab market share from its competitor, until then archidominant.

Added to this is Nvidia, which wants to switch to CPUs by trying to afford ARM. An Nvidia that rolls mechanics with its latest generations of Ampere chips and whose stock quote has surpassed that of Intel. Or Apple, which turns its back on x86 chips by developing its own ARM chips, which will ultimately deprive Intel of 7.6% of the IT market since it was Intel that supplied the brand with CPU. Here, too, Intel is “outdated”.

For all these reasons, Intel was probably a little fed up and also “dared” to compare its new baby to the competition. To try to remember that the number of cores or the frequencies are not everything.

More than a processor, a platform

In the dialectic of the presentation of Tiger Lake (11e generation of Core) one word is very important to Intel: platform. In addition to the assertions (we have not yet tested, ndr) of technical superiority over its direct competitor AMD in terms of pure performance, Intel intends to recall that it does not sell a simple CPU but an SoC (system on a chip , all-in-one chip) integrated into a complete platform, both software and hardware.

What is true: the EVO platform, which takes the fundamentals of Athena and which accompanies this 11e Core chip generation indeed contains more than a CPU / GPU pair. Wi-Fi (produced by Intel) is standard, as is Thunderbolt 4 (developed and again produced by Intel), AI acceleration is part (partly from Nervana) and software engineers from Intel infiltrates all major software developers to make sure their technologies are fully supported.

While AMD has the raw advantage in things like number of cores, Tiger Lake’s intrinsic performance is supported across the platform. The advantage of being a giant able to develop network chips, transfer protocols (Thunderbolt), etc. More than ever, this dominance of Intel highlights the quality and bravery of AMD engineers who have more than just a leader in front of them!

If we should not minimize the problems of production (loss of global leadership of its foundry), strategic governance (failure of mobile chips which left the field open to Qualcomm) or technological choices of Intel (without AMD, we it might still be quad-core!) it’s good to recall who Intel is. That is, number 1, which alone represents 17.6% of the overall semiconductor market (compared to 2% for AMD). Intel may be injured, but the tiger is still a tiger.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *