Refresh
Snapdragon X Elite and X Plus chips are potential game changers, but the entire computing world has been built around Intel’s x86 chips for decades. The entire ecosystem takes x86 as the default, so there’s no comparing Arm’s compatibility to what is already baked into the cake for Intel’s Lunar Lake.
I’ll have a lot to say about the SoC power efficiency. I haven’t been able to test any of it, but I’ve talked to a lot of Intel folks last week about Lunar Lake, and if the architecture changes Intel has made work as intended, we might be getting x86 performance with Arm-competitive energy efficiency. This chip has been entirely re-engineered, so a lot of the issues that devastated battery life from Alder Lake to Raptor Lake Refresh have been jettisoned.
Lion Cove performance cores, Skymont efficiency cores, Xe 2 (AKA Battlemage) architecture, and new NPU cores. Why he’s calling it ‘zhee-two’ is beyond me. Everyone at Intel has been calling it ‘echs-ee-two’, but when you’re the CEO, you can do whatever you want.
Intel Lunar Lake is official.
Intel Lunar Lake PCs on display.
OK, here we go, talking about AI PCs now.
Intel and Samsung bringing AI to edge devices.
UALink, the high-speed open interface standard for data centers that Dr. Lisa Su of AMD talked about yesterday, includes Intel, but noticeably doesn’t include Nvidia.
I have to admit, I’m glad they’re getting the data center stuff out of the way first. Interesting how AMD ended on data center and front loaded its client Ryzen announcements.
Inventec President Jack Tsai joins Gelsinger on the stage.
Multimodal RAG LLM analysizing an X-ray can help medical professionals interpret results for diagnoses. OK, so the ‘accidental’ reveal of this chatbot is a clever bit of presentation. I’ve watched Gelsinger do a few keynotes by now, and honestly, he’s better at this than anyone else in the industry. The content might not be the most riveting for most people, but he’s an excellent presenter.
Like I was yesteday at AMD’s keynote, I appreciate their focus on energy efficiency. There’s always the problem of induced demand, so these more efficient chips may just mean we use three times as many, so we end up spending the same amount of power (the same way adding more lanes to a highway doesn’t help traffic, since it just incentivizes more drivers to drive rather than take mass transit).
The difference between these two server racks is substantial when it comes to video. 4.2x performance gain on Xeon 6. Time to upgrade data centers off these older Intel Xeon chips. I agree. 80K MWh power savings with an upgrade to new Xeon 6 chips. I guess Intel read my oped.
LOL, a dig in at the guy’s workout routine. I mean, dude is in shape, so I guess he’s earned it.
That’s actually really good. Yes, move everything to e-cores please. I’m glad that the focus here is on efficiency.
Huh, Intel Xeon 6 with e-cores.
OK, getting data center talk out of the way. 130 million Xeon chips powering data centers.
Shout out to Netscape.
Chips are the heart of the global economy. Might not be such a great idea to have nearly all of them all being produced in a single country.
The AI Era is here, Gelsinger says, comparing it to the first days of the internet. He’s almost certainly right.
So that video of Intel founder Gordon Moore was generated by AI, with permission from Moore’s family. Which is fine, but I have to admit, I don’t like the idea of ‘resurrecting’ the dead in this manner.
Intel CEO Pat Gelsinger is about to take the stage.
So things are kicking off with a performance from the percussion group Ten Drum.
We’re about to be underway in about 4 minutes. Now’s a good time to grab a drink and a snack and settle in if you haven’t already.
Hey Folks! This is John, TechRadar’s components editor, and we’re about to start with Intel’s keynote address from Taipei, Taiwan. Expect a lot of AI PC talk from Intel’s CEO Pat Gelsinger, and much more besides.
+ There are no comments
Add yours