July 14, 2024
Watch All the things Introduced from AMD Keynote at Computex 2024 – Video

Watch Everything Announced from AMD Keynote at Computex 2024

Speaker 1: Welcome to everybody becoming a member of us as we speak in Taipei and from world wide as we open Compex 2024 at A MD. We love gaming. At the moment I am excited to indicate you what is subsequent for PC gaming with ryzen. Our new Ryzen 9,000 CPUs are the world’s quickest client PC processors bringing our new Zen 5 core to the AM 5 platform. With help for the newest IO and reminiscence applied sciences, together with PCIE 5 [00:00:30] and DDR 5, I am comfortable to indicate you now our model new Z 5 Core Zen 5 is definitely the subsequent large step in excessive efficiency CPUs. And if you take a look at the expertise behind this, we have now a lot new expertise. We’ve got a brand new parallel twin pipeline entrance finish, and what this does is it improves department prediction accuracy and reduces latency. It additionally permits us to ship far more efficiency for each clock cycle. We additionally designed Zen [00:01:00] 5 with a wider CPU engine and instruction window to run extra directions in parallel for management, compute, throughput and effectivity. Consequently, in comparison with Zen 4, we get double the instruction bandwidth, double the information bandwidth between the and floating level unit and double the AI efficiency with full AV X 5 12 throughput. All of this comes collectively within the product in RISE in 9,000 collection and we’re delivering a median of 16% larger IPC [00:01:30] throughout a broad vary of software benchmarks and video games in comparison with Zen 4.

Speaker 1: So now let me present you the highest of the road rising 9 99, 50 x for the very first time. There you go.

Speaker 1: We’ve got 16 Zen 5 cores, 32 threads as much as 5.67 gigahertz increase, [00:02:00] a big 80 megabyte cache at 170 watt TDP. That is the quickest client CPU on this planet. We all know all of our followers love gaming. The 99 50 x delivers finest in school gaming efficiency throughout a variety of fashionable video games. Now with desktops, we all know that lovers wish to have an infrastructure that allow you to improve throughout a number of product generations. And we’re with Verizon. We have executed simply [00:02:30] that. Our unique Ryzen platform socket AM 4 launched in 2016 and now we’re approaching the ninth 12 months and we have now 145 CPUs and apus throughout 11 totally different product households in Socket AM 4. And we’re truly nonetheless launching new merchandise. We even have just a few ryzen 5,000 CPUs which can be coming subsequent month, and we’re taking this very same technique with Socket AM 5, which we now plan on supporting by 2027 and past. [00:03:00] So you are going to see AM 5 processors from us for a lot of, a few years to come back. Now along with the highest of the stack rising 99 50 x, we’re additionally saying the 12, eight and 6 core variations that can convey the management efficiency of Zen 5 to mainstream worth factors and all of those go on sale in July.

Speaker 1: We’re so excited to announce as we speak our third gen rise in AI processors. Our information rise in AI [00:03:30] collection truly is a major improve in compute and AI efficiency and units the bar for what a copilot PC ought to do. Thanks, drew. Right here we go. That is STRs. strix is our subsequent technology processor for extremely skinny and premium notebooks, and it combines our new N 5 CPU sooner RDNA 3.5 [00:04:00] graphics and the brand new XD NNA two NPU. Thanks.

Speaker 1: And if you take a look at what we have now, it truly is all the finest expertise on one chip. We’ve got a brand new NPU that delivers an business main 50 tops. We’ll speak about tops so much as we speak. 50 tops of compute that may energy new AI experiences at very low energy. We’ve got our new N 5 core [00:04:30] that allows all the compute efficiency for ULTRATHIN notebooks and we have now sooner RDNA 3.5 graphics that basically brings the most effective in school software acceleration in addition to console stage gaming to notebooks. Now we have now a few skews. The flagship risen 9, HX three 70 has 12 and 5 cores, 24 threads, 36 megabytes of money. The business’s most built-in NPU and our newest RDNA graphics. strix is just the most effective [00:05:00] cellular CPU. So lemme discuss a little bit bit about what’s particular on this new NPUN PS are actually new they usually’re actually there for all of those AI purposes and workloads.

Speaker 1: Now in comparison with our prior technology XD A two options a big array of 32 AI tiles with double the multitasking efficiency. It is also a particularly environment friendly structure that delivers as much as two instances higher power effectivity of our prior technology when operating [00:05:30] gen AI workloads. Along with Microsoft, we’re additionally working with all the main software program builders, together with Adobe, epic Video games, SolidWorks, Sony, zoom, and plenty of others to speed up the adoption of AI enabled PC apps. And by the top of 2024, we’re on monitor to have greater than 150 ISVs growing for a MD AI platforms throughout content material creation, client gaming and productiveness purposes. Now to provide [00:06:00] us a take a look at a few of these upcoming copilot plus PCs, let me welcome our subsequent visitor, a really shut associate and good good friend Enrique Lores, HP president and CEO. Hi there. Hi there.

Speaker 2: And we’re extremely excited concerning the new household of merchandise we might be launching in just a few weeks as a result of

Speaker 1: I feel you’ve gotten one thing to indicate us. Is that proper? Okay. I hope so.

Speaker 2: And that is truly the brand new [00:06:30] technology since we have now executed collectively, we will present it collectively. That is great. That is the subsequent technology omni guide after we combine the newest current AI 300 collection that, as Lisa was saying earlier than, would be the first product that can have 50 tops built-in within the system. And efficiency, as Pavin was saying, is crucial as a result of will [00:07:00] allow us to proceed to ship unimaginable experiences to our prospects.

Speaker 1: Let’s welcome Luca Rossi, president of Lenovo Clever Units Group. Now Luca, you have been holding one thing and I feel you are going to present us what it’s. Is that proper?

Speaker 3: Nicely, I wasn’t imagined to even present this but since that is one thing we won’t announce till later this 12 months, however Lisa, you might be proper. I felt it’s simply too thrilling [00:07:30] to you look

Speaker 1: Very comfortable truly.

Speaker 3: Yeah, very comfortable. So straight from our r and d, straight from our r and d lab. That is the primary ever sneak peek at our new yoga laptop computer powered by third gen rising ai. That is

Speaker 1: Lovely.

Speaker 3: Possibly we wish to do left and proper. Yeah, so I [00:08:00] can share extra for as we speak. I can can. That is very lovely. Thanks. I can let you know this system symbolize a major leap ahead in subsequent technology AI computing will embody a number of the Lenovo unique IP AI options that I discussed and that is just the start.

Speaker 1: Subsequent, I might wish to welcome one of the crucial necessary visionaries and innovators within the Taiwan ecosystem and a really, very shut associate Johnny she chairman of Jesus.

Speaker 4: [00:08:30] I feel later as we speak at 4:00 PM we’ll be unveiling a spread of innovative IPCs throughout our portfolio with model new Zen Guide professional artwork. We guide a stuff and IG laptops powered by the third gen A MD rising AI processors. These new land labs are outfitted with the world’s strongest MPU with 50 tops [00:09:00] and the superior MB xFi structure that leads the business in compute and AI efficiency. The third gen rising AI processor is the catalyst to herald customized computing to everybody from content material creators to avid gamers and enterprise professionals and empowering them like by no means earlier than. This development give the brand new handbook larger AI efficiency than MacBook [00:09:30] whereas making it thinner and lighter as effectively. Truly is so proud and honored to be the primary OEM associate to make the third gen rising AI programs out there to customers. It is going to be prepared for buy in July. Is not that unimaginable? Thanks.

Speaker 1: So I’ve one thing to indicate you as we speak. It is truly the preview of our upcoming fifth gen Epic processor code named flip.

Speaker 1: [00:10:00] So please check out flip for the very first time flip. Options 192 cores and 384 threads and has 13 totally different chips inbuilt three and 6 nanometer course of expertise. There’s a whole lot of expertise on Turin. It helps all the newest reminiscence and IO requirements and is a drop in alternative for our present [00:10:30] fourth gen Epic platforms Flip will lengthen Epic’s management normally goal and excessive efficiency computing workloads. So let’s check out a few of that efficiency. namd is a really compute intensive scientific software program that simulates advanced molecular programs and constructions. When simulating a 20 million at a mannequin, A 128 core model of Turin is greater than 3 times sooner than the competitors’s finest, enabling researchers to extra rapidly full fashions [00:11:00] that may result in breakthroughs in drug analysis, materials science, and different fields. Now Turin additionally excels at AI inferencing efficiency when operating smaller giant language fashions.

Speaker 1: So I wish to present you a demo right here. Now. What this demo compares is the efficiency of Turin when operating a typical enterprise deployment of LAMA two digital assistant with a minimal assure latency to make sure a top quality person expertise. Each servers start by loading a number of LAMA two cases [00:11:30] with every assistant being requested to summarize an add a doc instantly you may see that the flip server on the precise is including double the variety of periods in the identical period of time whereas responding to person requests considerably sooner than the competitors. And whereas the opposite server reaches a most variety of periods, you may see it cease quickly. It principally cannot help the latency necessities anymore. Flip continues scaling and delivers a sustained throughput [00:12:00] of almost 4 instances extra tokens per second. Our prospects are tremendous enthusiastic about flip and I do know a lot of our companions are literally right here within the viewers as we speak. And I wish to say we’re on monitor to launch within the second half of this 12 months. Thanks for being such an important viewers and have an important Compex 2024. Thanks.

Speaker 5: Thank.