As a part of the ultimate phase of Intel’s “Intel Consumer Open Home Keynote” at CES this afternoon, Intel EVP and GM of the Consumer Computing Group, Michelle Johnston Holthaus, additionally provided a short replace on Intel’s shopper chips within the work for the second half of the yr. Whereas no demos had been run through the comparatively brief 45 minute keynote, Holthaus did reiterate that each Arrow Lake for desktops and Lunar Lake for cellular had been making good progress and had been anticipated to launch later this yr.
However in lieu of black field demos we obtained one thing extra shocking as an alternative: our first have a look at a completed Lunar Lake chip.
Briefly holding the chip out for viewers to see – and holding the press away least they get too shut – Holthaus pulled out a completed Lunar Lake chip.
Whereas particulars on Lunar Lake nonetheless stay very slim – Intel nonetheless hasn’t even confirmed what course of nodes it’s utilizing – the corporate has regularly been reiterating that they intend to get it out the door in 2024. And having silicon to point out off (and delivery to companions, we’re informed) is a really efficient solution to exhibit Intel’s ongoing progress.
Of observe, seeing the chip in particular person confirms one thing we’ve all however been anticipating from Intel for a couple of years now: CPUs with on-package reminiscence. The demo chip has two DRAM packages on one of many edges of the chip (presumably LPDDR5X), making this the primary time that Intel has positioned common DRAM on a Core chip bundle. On-package reminiscence is of explicit curiosity to skinny & mild laptop computer distributors, because it permits for additional area financial savings and cuts down on the variety of crucial traces that have to be routed from the CPU and alongside the motherboard. The method is most notably (although removed from solely) used with Apple’s M sequence of SoCs.
Past exhibiting off the bodily chip, Holthaus additionally briefly talked about anticipated efficiency and structure. Lunar Lake is slated to supply “important” IPC enhancements for the CPU core. In the meantime the GPU and NPU will every supply three-times the AI efficiency. How Intel can be reaching this stays unclear, however at the least on the GPU facet, we all know that they’ve but to supply XMX matrix cores inside an built-in GPU.
Little question that is removed from the final time we’ll hear about Lunar Lake forward of its launch. However for now, it’s a little bit of a glance into the longer term whereas Intel continues to ramp manufacturing on Meteor Lake for what’s now the most recent technology of laptops and different cellular gadgets.
