Watch Intel Launch Their New Core Ultra Processor at IFA 2024 video [CNET]
Welcome to Intel’s next generation core ultra global launch. Today, we’re proud to announce our newest core ultra processor, codenamed Lunar Lake. Lunar Lake has significant architectural innovations, cores, graphics and A I. It’s the most power efficient processor we have ever built. We’ve overhauled the entire architecture, its design and its construction. First, the performance core, it’s wider, faster and more scalable than any core we’ve ever built. It’s the new foundation for how we think about core design. Moving forward compared to our last P core, it delivers double digit increases in performance performance per watt and performance per area with Lion Cove. Lunar Lake delivers the fastest CPU core of any processor in a thin and light laptop. Let’s move on to the second type of core. The efficiency core, we code name Sky mot. It has two X the throughput. Pray I it has a much better and enhanced branch predictor. It’s wider and deeper for more throughput and has a massive 68% I PC increase. Yes, 68%. Now let’s move on to graphics. The overall in our graphics processor is just as remarkable. Lunar Lake is the first product to use our new XE two graphics architecture and Xe two delivers substantial improvements in energy efficiency. There are eight graphics cores in Lunar Lake, the average performance lift of each core about 30% in many games. Example, Cyberpunk, Spiderman Hogwarts legacy, we’re seeing performance improvements up to 50%. The crowning achievement of ex two is simple. It’s the fastest built in GP U in the industry to prove that we’ve tested nearly 50 different games and we’re faster than our competitors best 16% faster than AMD, 68% faster than Qualcomm. And that was actually very difficult to measure because 23 games didn’t run on the arm processor. Gaming is more than just performance. Our investment in drivers and day zero game compatibility are a huge part of the experience for our customers. Now, we’d like to show a demonstration of the R GP U in action running on Lunar Lake. All systems are running the same game dota two. We’re going to look at performance and power. So first up, look at Lunar Lake versus DA MD on the left. Both are running about the same frame rates, but Lunar Lake is running at about 30% lower power. Then if you move to the right hand, right hand side, you’ll see Lunar Lake versus Qualcomm and Lunar Lakes running about 25% faster. But at half the power, if we do the math, lal la laptop delivers 35% more energy efficiency than A MD and over twice that of qual at the end of the day. However, we know performance is king for everyone watching. So let’s recap. We’re 30% faster than our prior gen 16% faster than AMD and 68% faster than Qualcomm. Now, let’s move on to A I. We have 100 and 20 total platform tops in Lunar Lake and every one of our engines have been increased in their capability. The NPU is four times bigger. The GP U has added A I extensions and we’ve doubled the A I throughput of the course XE two has another trick called XE matrix extensions. We call it XMX which delivers up to 67 tops just from the GP U. The Lunar leg GP U is the fastest A I accelerator in a mobile processor ever building on the GP U. We also have a new NPU nearly four times the performance of our prior gen up to 48 tops. Other MP us have similar specs on paper but we deliver the best performance by doing more with every top because our developers are focused on optimizing their applications and enabling new workloads versus fixing legacy compatibility issues. Now, let’s talk about power efficiency. Lunar Lake systems shatter the myth that X 86 can’t do low power and long battery life. Lunar Lake is the most efficient X 86 processor ever. The efficiency in Lunar Lake is far more than co it’s packaging. It’s the fabrics. It’s how we cash and of course, process technology compared to our previous generation, we slashed soc power 50% when it came to battery life. We pulled out all the stops. Laptops based on Lunar Lake can deliver 20 hours of active and productive battery life. In fact, we measured it this week on a customer system ready to launch and that’s 20 hours in a workload that is much more demanding than simply watching a video on a dimmer screen. Let’s finally touch on compatibility before you can even experience performance or power. You need your application to just run. There’s nothing more frustrating than on your app. When you load your app, it doesn’t work. Core alter is built on the proven platform of X 86 with software compatibility. You don’t need a website to check if your favorite app will run. Generational compatibility is one of the P CS greatest strengths and L Lake does not break that trust. And in this spirit, we’re expanding our A I efforts to another operating system. Please join me in welcoming John Solomon from Google. Thank you John for coming to Berlin. So rather than me, just talk about it theoretically, um Let me show you some examples. So um you can look on the screen behind me. But what we, what you’re seeing there is help me read. This is a new A I powered feature we’re introducing. Um actually later this month, September already, right? Yeah, thanks for showing it here for. Yeah, it’s the first time we’re actually showing it publicly so we can all applaud. And you know, this is a, this is a, this is a kind of core productivity feature that we’re focused on. Um you know, this is something we all do in our, in our real world. In fact, um if you think about help me read in action, the Intel team sent me a long detailed email, actually multiple. I’m sorry to help you prepare for this event. And like most of you, I’m running late. I’m a little behind on my emails. So I’d have as much time to, to read all my emails. So I used help me read to figure out what the heck I was supposed to do when and just incredible productivity saver. So I was super happy about that. Um You’ll also see uh in the in the gift behind me here, Magic editor. This is one of my personal favorites where you can quickly make advanced edits to photos with A I. Uh whether that’s un blurring old photos of yourself or your friends or in the case that you’ll see on the, on the slide. Um I needed to make myself a little bigger in the, in the, the image. So I took advantage of that. So be careful when you look at a slide, you never know exactly what that’s, I know you’re a cloud client. That’s, that’s how you roll. But this is running on device and it means it’s lightning fast. It’s local and it’s mine and I can work while I’m on the move Intel has great hardware to run A I locally on your PC. But it’s not always easy for everyone to get started to give everyone the opportunity to try A I on their PC. Intel has released an open source project called A I Playground built specifically for Intel core ultra and art graphics A I playground is a simple to use Windows app with a really user friendly interface. You can get started with just one click. It’s free, it runs entirely locally and has all of the latest capabilities A I playground allows you to select from a rich set of A I models, right sized for local execution on your Intel PC. So let’s check it out with Lee stable diffusion is a gen a a model that allows you to generate images from text or even other images. Here, Lee is prompting the model with Lunar and Lake. Of course to create these beautiful photos you see right here. The inference runs almost instantly on the R GP U fully local no cloud needed from here. We can switch right to a small language model using what’s called retrieval augmented generation or rag. Rag allows you to directly introduce new skills or knowledge to personalize your A I. It’s like having a personal chatbot running your data locally on your Intel PC. And here we see a great example for developers generating Python A I playground is a great example of how we think about A I at Intel. It’s open, accessible and secure. It’s a great way to get started if this is your first time using A I on A PC. And it’s also a great way to get a jumpstart. If you’re a more advanced user, we hope you love it on your new core ultra laptop. So if that wasn’t enough, I have one more thing to share pre orders for Lunar Lake start today. Exactly. And you can see systems on retail shelves starting September 24th. We have more than 80 unique designs across more than 20 top O ems many of our partners here in the audience being sold in more than 30 retailers around the world. Only Intel can deliver this scale. What you see over here on your right. And my left is a device wall of high quality premium systems from all of our top OEM partners. And we can’t wait for each of you to get your hands on one.