So Intel has lifted the lid on its underlying Lunar Lake mobile CPU, and it's very much absorbing. It is entirely new. A new CPU core, a new graphics architecture, a new approach to both the internal fabric of the chip and its multi-die structure, a new external Fa and, generally true to Intel's current mojo, Lake of the Moon is a fascinating mix of wonderful, strange, worrisome and, well, nothing.

The great bit is the new E-core that Intel says is almost equivalent to the full P-core of today's Raptor Lake desktop CPUs for performance per clock To get full performance from these new Skymont E-cores, it needs to be fully integrated into the CPU's high-performance fabric, but this is not the case. This new process

in Lunar Lake, it is isolated into so-called low-power islands in the name of efficiency, so it has less cache than other methods and less optimal access to the available high-level cache. But you're still seeing a big step forward in the E-core of other Intel CPUs.

Indeed, Intel believes that lunar Lake is designed to complete almost everything while running on the battery power of those E-cores. If ever get fired while on the battery, the P-core is rarely.

Lunar Lake's P-core, known as Lion Cove, also gives a healthy boost to the ipc. As a result, Intel claims that the Lunar Lake processor with 4 E-cores and 4 P-cores is better than Intel's current Meteor Lake laptop CPU with 10 E-cores and 6 P-cores.

Put another way, it's the 22 moon lake threads that hit the 8 meteor Lake threads. The thread count gap is partly down to Intel dropping the hyperthreading of Lunar Lake's P-core. Impressive, but you see it.

Another potentially great element is the new Battlemage iGPU, which will take over eight running units from Meteor Lake, but still has about 50% performance thanks to a massive redesign.

Alternatively, Lunar Lake can be matched with Meteor Lake's game performance, and the power used is about 35%. 〜 40% less. Taking it all together, Lunar Lake has the makings of one heck of a handheld gaming PC chip. If anything, it looks more promising than amd's new Strix Point, which looks faster, but worries about power consumption.

In addition, Intel is definitely putting more effort into this chip to reduce power consumption than ever before.The E-core is completely separated from the P-core, and the internal fabric and layout of the chip is fundamentally different from previous Intel designs, the number of tiles has been reduced in Meteor Lake, and the memory has been transferred to the package. Well, you get the idea.

So this is Intel's best effort. It is also a good omen for the future. We can't wait to see these new cores on the desktop, and the same applies to the Battlemage GPU architecture. It can't come soon enough in discrete desktop format.

For something strange and worrisome, let's start with manufacturing technology. Lunar Lake's compute tile is based on TSMC N3B, but it's not very positive about Intel's internal process nodes.

Shortly after he returned to Intel as CEO, Pat Gelsinger set what he claimed was a bold goal for 4 new nodes from Intel in 5 years. Lunar Lake has not yet been on sale for several months, and somehow there are not enough Intel nodes for this chip.

The choice of N3B itself is also definitely a bit odd. SO FAR, ONLY APPLE IS USING THE N3B NODE IN TSMC. The yield and cost of N3B is said to be pretty cheesy, so everyone else is waiting for a cheaper and better yielding N3E node from TSMC.

Indeed, the latest scuttlebutt is that Apple can't get out of the n3b fast enough, so the new M4 chip in the n3E looks like a killer on the N3B-based M3 si N3E, but the n3b-based Apple M3 chip has a relatively speaking performance advantage. I was disappointed for the performance and efficiency. To be fair, N3B has not been TSMC's best time.

Lunar Lake's timing and market positioning also feel a bit irregular. Meteor Lake family of laptop chips really only became available in the right volume earlier this year, and already Intel is rolling out a whole new laptop chip.

Of course, Lunar Lake is not a complete family of laptop chips, it just covers a low-power niche. The problem that Moon Lake is fixing may be Meteor Lake. But if Intel provides a roadmap, there will be yet another new mobile CPU family in Panther Lake by the end of next year. But you slice it and it all looks a bit ad hoc and messy.

It's also a bit worrisome that Intel showed no detailed performance numbers or signs of clock speed on the new chip. True, this is not a product launch, so the final specifications may not be much available. But the hard data was pretty thin on the ground.

If I had to guess, I would say Lunar Lake is rushing for 2 reasons. First, Meteor Lake may not have been as effective and efficient as Intel had planned. It was certainly not a big step forward for performance or battery life. Second, Meteor Lake is not subject to Microsoft's Copilot+AI PC definition because Npu isn't powerful enough. Intel didn't want to wait until later next year for Panther Lake to get a co-pilot+AI product to compete with Amd's Ryzen AI300 chip and Qualcomm's Snapdragon X Elite. That's why Intel desperately needed a chip with a more powerful NPU.

And we have the Moon Lake in all its wonderful oddities. Honestly, I don't know if the design is a sign that innovation is still alive at Intel, or if it's evidence that the house isn't in place yet. Enough is going on to make a half-decent case for both discussions. But no matter what you think about Intel's new mobile chip, it's certainly not boring.