AMD Unveils Ryzen AI 300 CPUs Technology Supercharging AI on Laptops

AMD Unveils Ryzen AI 300 CPUs Technology Supercharging AI on Laptops

AMD has unveiled its next generation of Ryzen processors, with the Ryzen AI 300 series taking center stage for laptops. Announced at Computex 2024, these CPUs are designed to usher in a new era of AI-powered experiences on portable machines.

The key feature of the Ryzen AI 300 is its all-new Neural Processing Unit (NPU). This built-in chip boasts a whopping 50 TOPS (tera operations per second) of performance, which is more than three times what AMD offered in its previous generation Ryzen AI processors. This leap positions AMD as a leader in on-device AI processing for laptops, directly addressing the needs of Microsoft's Copilot+ initiative.

Copilot+ aims to create a new standard for AI-powered PCs. By combining powerful CPUs, GPUs, and NPUs, Copilot+ laptops can handle demanding AI tasks that were previously limited to desktops or the cloud. The Ryzen AI 300 CPUs technology, with its best-in-class NPU, is perfectly suited for these next-generation AI laptops.

But the Ryzen AI 300 offers more than just AI muscle. The CPUs are built on AMD's new "Zen 5" architecture, delivering significant performance gains over previous generations. This includes up to 12 high-performance cores with 24 threads, offering excellent multitasking capabilities. Additionally, the Ryzen AI 300 boasts a 50% increase in on-chip L3 cache memory compared to Zen 4 processors, improving overall responsiveness for everyday tasks.

While the focus is on AI performance, gamers will also be interested. The Ryzen AI 300 CPUs are designed to work seamlessly with powerful graphics cards, making them strong contenders for mobile gaming laptops.

Several major manufacturers, including Acer, ASUS, HP, Lenovo, and MSI, have already announced their intention to incorporate the Ryzen AI 300 series into their upcoming Copilot+ compatible laptops. This means that consumers will soon have a wide range of choices for AI-powered laptops, from thin and light machines to gaming powerhouses.

The launch of the Ryzen AI 300 series marks a significant step forward in bringing powerful AI capabilities to laptops. With AMD's latest innovation, users can expect a new level of performance and intelligence on their portable devices

• Supercharged AI: AMD unveils the Ryzen AI 300 series CPUs for laptops, featuring a best-in-class NPU with 50 TOPS performance – over 3x faster than previous Ryzen AI processors.
• Copilot+ Ready: Perfectly suited for Microsoft's Copilot+ initiative, enabling demanding AI tasks directly on laptops.
• Zen 5 Powerhouse: Built on AMD's latest Zen 5 architecture, offering significant performance gains with up to 12 cores, 24 threads, and 50% more L3 cache for smoother multitasking.
• Gamers Rejoice: Optimized for powerful graphics cards, making Ryzen AI 300 a strong contender for high-performance mobile gaming laptops with latest technology.
• Widespread Adoption: Major manufacturers like Acer, ASUS, HP, Lenovo, and MSI are incorporating Ryzen AI 300 into their upcoming Copilot+ laptops, offering a wide range of choices for consumers.

Conclusion:

The Ryzen AI 300 series marks a turning point for AI on laptops. Users can expect a significant boost in performance and intelligence for tasks ranging from content creation to cutting-edge AI applications.

Apple plans to move to graphics AMD?

Reporters found on Apple's website a new job. Company is looking for an engineer in the field of hardware with the experience of working with graphics solutions from AMD or NVIDIA. The priority is not given to any manufacturer.

Who is the supplier of graphics processors for Apple computers is NVIDIA. Both parties appear to be satisfied with the cooperation, so this news is not a reason to direct proof intentions Apple again adopt a schedule AMD. Meanwhile, the theme provides a fertile ground for all kinds of speculation.

According to the description jobs, the new Apple employee will develop electronic systems for candy bars iMac. In addition to knowledge in the field of graphic solutions, the candidate must understand the chipsets Intel (Apple is categorical) and DDR3 memory interface and GDDR5.

By the way, the story has it that in the past year, Apple has been one step from being used in a new line of ultra-thin MacBook Air laptop crystal AMD LIano - the so-called APU (Accelerated Processing Unit, or "accelerated processing unit") - graphics adapter, memory controller and video decoder on one chip.

Then computing solutions from AMD fully meet the needs of Apple, but the deal fell through because of the chip maker. The company would be unable to ensure the supply of adequate number of processors due to its transition to production partner Global Foundries to new technological standards, and there was not enough as the final sample.

Represented by the world's fastest single-processor 3D-card - AMD Radeon HD 7970 GHz Edition

As expected , AMD has released a 3D-card AMD Radeon HD 7970 GHz Edition. According to the company, the new product is the world's fastest single-processor 3D-card. Thus, AMD regained title of the world's fastest single-processor 3D-card, was briefly intercepted by NVIDIA.

 

C a technical standpoint AMD Radeon HD 7970 GHz Edition differs from the AMD Radeon HD 7970 , AMD is in December last year opened an era of 28-nanometer GPU, only the high frequencies components. Clock speed GPU Tahiti with 2048 stream processors is equal to 1050 MHz (in the case of AMD Radeon HD 7970 - 950 MHz) RAM - 1500 MHz (AMD Radeon HD 7970 - 1375 MHz).

In the configuration of the product includes 3 GB of memory GDDR5, which are connected with the GPU 384-bit bus. It features proprietary technologies AMD Eyefinity 2.0, App Acceleration and CrossFire. It is equipped with interface PCI Express 3.0.

Sales of AMD Radeon HD 7970 GHz Edition partners AMD, including ASUS, Gigabyte, HIS, MSI, Sapphire, TUL and XFX will begin next week. Price new - $ 499.

MSI R7870 Hawk makes the graphics card

MSI enriches the catalog of products in the Hawk , for enthusiasts and overclockers, featuring graphics card R7870 Hawk , which inherits many features from the model R7970 Lighting. The MSI R7870 Hawk measuring 277 x 120 x 405 mm, it has 2GB of GDDR5 memory and has operating frequencies for GPU and memory respectively 1100MHz - 1200MHz (4800MHz effective).

The card uses the heatsink Twin Frozr IV equipped with two fans eighth mm; the radiant body is connected to the GPU and to the components of the PCB by means of two copper heatpipes 8 mm and three 6 mm. According rirefisce the manufacturer, this heatsink lowers operating temperatures of up to 15 ℃ compared to stock coolers, also enhancing the noise.

Going deeper into the technical features, the Radeon HD 7870 from MSI proposal presents well 8 + 2 phase power that ensures stability under overclocking important. Finally we point out the composite video output to a DVI port, one HDMI and two DisplayPort. No mention about the official date of its commercialization of the sale price.

Processors AMD FX-8150 in the European retail considerably cheaper

Processors AMD FX-8150 can compete with the products of generations of Intel Sandy Bridge, such as the Core i5-2500K, but the output of the Quad desktop processors Intel Ivy Bridge broke the balance in this segment. To remedy the situation and improve the relationship between price and performance of its processors, AMD has slashed prices on the FX-8150 on the European market. As the data sources, price FX-8150 is now starting to EUR 179 (including VAT). Recall, after the October decline , this model has cost 213 euros.

 

Interestingly, the manufacturer changed the packaging of the product. If before the processor AMD FX-8150 was placed in a tin box (top illustration), but now it has given way to more familiar to buyers of a cardboard box (lower illustration).

Obviously, the transition to a cheaper package - one of the reserves to reduce price.

AMD is going to lower the price of 3D-card Radeon HD 7970, HD 7950 and HD 7700

AMD did not rush to lower prices for its 3D-maps of a 28-nanometer GPU in response to the release of the first model, the new generation by NVIDIA. Perhaps the superiority of NVIDIA GeForce GTX 680 on AMD Radeon HD 7970 AMD does not seem too convincing. Perhaps the volume of deliveries 28-nanometer production lags behind demand. Somehow, over time, AMD will still need to go for lower prices in order to strengthen its position before the release of the following models the architecture of NVIDIA's Kepler and support the demand, when no longer affect the novelty effect.

According to sources, AMD is already preparing the first stage to reduce the price of 3D-card series Radeon HD 7900. As stated, the new recommended price model of Radeon HD 7970 will equal $ 489, Radeon HD 7950 - $ 394. This means that the older model cheaper at $ 60 (now its recommended price is $ 549), and the youngest - to $ 55 (it is now worth $ 449). From a psychological point of view, the price will go to another step of perception - with "more than $ 500" to "$ 500" and "more than $ 400" to "$ 400".

And cheaper model Radeon HD 7770. But not so much in absolute terms - for $ 15.

When the new prices will be announced, the source is not reported.

TSMC is having difficulty with the release of 28-nanometer production

According to industry sources, the world's largest contract manufacturer of semiconductor products TSMC is having difficulty meeting demand for 28-nanometer products. The main customers are companies TSMC Qualcomm, AMD and NVIDIA. It is believed that the situation will improve in the third quarter of this year.

However, Qualcomm agreed to transfer part of their orders for ARM processors for the company United Microelectronics (UMC), because the volume of products manufactured by TSMC, is not enough for the rapid market of smartphones and tablets.

AMD has introduced the first graphics accelerator based on a 28-nanometer GPU at the end of last year, but shipments family of video cards Radeon HD 7000 is still quite small.

NVIDIA has managed to release two cards based on 28-nanometer GPU Kepler: GeForce GTX 680 desktop PCs and GeForce GT 640M mobile, but, apparently, will also be forced to postpone plans.

For this reason, TSMC is considering the expansion of production . Construction of a new line in the factory Fab 14, which works with 12-inch (300 mm) plates, will begin April 9. Total for the year 2012 the company plans to increase production by 10%.

Issue APU AMD Trinity is scheduled for May 15

While Intel processors occupy first place in the niche of high-performance processors, AMD has an opportunity to compete for leadership in the middle class.

On the slide, AMD indicated that the new APU Trinity is 29% better performance and a 56% superior to the graphics processor of the APU Llano. Also there is information regarding the optimization of APU Trinity for the future of Windows 8.

The new processors will support a more advanced version of the dynamic overclocking technology - Turbo Core 3.0. A mobile version processors allow more efficient use of battery power, which enables mobile devices to work over 12 hours.

There will also be HD Media Accelerator - a set of algorithms that will improve the image quality.

According to sources, the first model of the family will come May 15, 2012. by 

Dell Alienware M18X R2 with Dual-GPU setup?

Although not yet officially announced by Dell , the new notebook Alienware M18X R2should be equipped with the latest generation graphics cards Nvidia or AMD Dual-GPU solutions probably double or tab. There is talk of GeForce GTX 660M and GTX 675m(the latter will be available in SLI configuration), while the AMD Radeon HD 7970M will be configured in CrossFireX.

In addition to this the new Alienware M18X R2 will have a 18.4-inch screen, a processor Intel Ivy Bridge, a black or red body board and wireless HD with a wide scope for expansion. Unfortunately, Dell did not provide further details.

Published results APU AMD Trinity A6 for ultrathin notebooks in 3DMark

Speaking of hybrid CPU AMD Trinity A6, designed for ultra-thin notebooks (Platform Ultrathin), AMD has published the result, which shows that the CPU test in 3DMark.

The result is 2355 points. For comparison - the processor Intel Core i5-2537M (Sandy Bridge) is gaining in the same test score of 1158. Both products are classified as low-power processors - the TDP value of each of them is 17 Tues

At AMD, suggest that the performance of Intel processors with the micro-Ivy Bridge will be 30% higher than the performance of Intel microarchitecture Sandy Bridge. Accordingly, the expected result can be Ivy Bridge in the test 3DMark - 1505. In other words, a Trinity, surpassing its current competitor by 103%, and may retain the superiority of his successor - by 56%.

Model APU AMD Trinity A10 with a TDP of 25 watts, designed for a "thick" notebook shows the result in 3DMark 3600 scores. This is 136% greater than the expected figure Ivy Bridge.

Specify that it was a higher graphics performance. As for the performance of CPU, it is expected that the product Intel will have the best performance based on a single core and single-threaded applications. In multithreaded applications, the position of Trinity stronger, because APU with two modules (actually - four cores) will be able to compete on price with the dual-core Intel, able to carry four streams through technology Hyper-Threading.

Interestingly, the presentation of test results has been removed from public view.

Six-core AMD FX-6200 officially approved in the range of AMD

About two weeks ago , we reported that AMD is preparing to release a new processor architecture for Bulldozer, FX-6200. Now this model has a range of companies listed on the official site, ie, its existence was confirmed on the official level.

As reported earlier, AMD FX-6200 is equipped with six cores, the nominal clock frequency of 3.8 GHz. With the support of technology Turbo Core 2.0 CPU reaches a maximum frequency of 4.1 GHz.

The technology, like its predecessor AMD FX-6100, produced at rates of 32-nanometer manufacturing process, has an unlocked multiplier and is equipped with 6 MB cache second level, and 8 MB cache in the third level. Unfortunately, the creators could not keep the TDP FX-6200 FX-6100 levels: 125 W vs. 95 W .

Approximate price of AMD FX-6200 - $ 175 U.S. and 150 euros in Europe, making it slightly cheaper than new product direct competitor - Intel Core i5-2400. At the same time, AMD FX-6200 will cost $ 20 more expensive than its predecessor: the price of AMD FX-6100 in the U.S. - $ 155.

AMD Radeon HD7990: a gigantic with 6GB of memory GDDR5

After the recent presentation of the Radeon HD7970 (you can read Review: AMD Radeon HD7970, raw power to 28Nm ) and imines dell'HD7950 debut, AMD is also preparing to launch dual-GPU HD7990 solution. The "queen" of the video cards on the market in Q1 and will be a monster of power as it will be equipped with 4096 shaders, 62 Compute Units and 6 GB of GDDR5 video memory (3GB 2x) with dual 384-bit bus.

Given the power saving features of the HD7900 series, the frequencies could be the same as the Radeon HD7970, then most likely we will be two GPU to 925 MHz to 5.5 GHz with Memory The price, however, will not be quite popular.

This is the first Asus and HP with 28-nanometer GPU

The first 28 nanometer AMD GPUs have appeared in the specifications of the HP and Asus notebooks; the GPU in question are the Radeon HD7670M and HD7690M, both based on the chip architecture WLIW Thames XT-4 and 128-bit GDDR5 memory.

HP has also listed a Radeon drivers HD7470M on its website, all in conjunction with an event held in London on their own Radeon HD 7000M. The notebook HP Envy 17-3030ew will then have a Radeon with 1GB GDDR5 HD7690M along with a processor Intel Core i7 SandyBridge.

The ' Asus X53TK SX058V- hand is a notebook with Radeon HD 7670M in addition to a platform with Llano APU. AMD is expected to announce officially to 28Nm video cards for laptops directly to the event taking place in London in recent days.

AMD Bulldozer FX: Good sales, stops the production of the Phenom II

Despite the enthusiast user does not have appreciated fully the absolute performance of the new AMD FX (you can read  Review:AMD FX-8150, 8 Core 3.6 GHz ), sales of the first units arrived in the shop and the stocks OEMs have clocked good numbers, facts already running out of stocks available.

AMD has so much steel and copper alloys of a request processor architecture Bulldozer FX in addition to that of the Llano chip platform. This strong demand seems to have brought home to Sunnyvale to stop the production of the Phenom II and Athlon II.

This move, taken from the site Nordichardware, may not be well accepted by users who look at absolute performance, in particular also because we remember that the Phenom II X6 gave a hard time with the new FX especially in multi-threaded applications.

In short, a choice that can be criticized but that in a market perspective it seems quite logical as the AMD FX high-end models allow the manufacturer a much higher gain.

The next year, AMD will introduce chipsets 1090FX and 1070 processors in the performance of AM3 +

Despite the fact that the chipset 990FX and 970, on which are based motherboards for processors in the performance of AM3 +, not even half a year , AMD is already working on their successors. So, next year at their 990FX and 970 chipsets will come 1090FX and 1070 respectively, the first details of which have become known recently.

The most interesting feature of the AMD 1090FX, in comparison with its predecessor, will support USB 3.0 natively, rather than through third-party controllers. Unfortunately, the number of ports, USB 3.0, implemented by means of the chipset, not reported, but given that the A75 provides four such connectors, it is hoped that AMD 1090FX provide work at least as a smaller number of ports.

Another interesting feature of AMD 1090FX - Support for up to eight ports SATA 6 Gb / sec. For comparison, the Intel X79 chipset provides the only two ports SATA revision 3.0. Finally, the new chipset will support AMD 32 lines of interface PCI Express, which will be divided equally between the two slots for video cards. Supporting AMD CrossFireX technology and Nvidia's SLI, of course, also be present.

As for the chipset 1070, he, unfortunately, will not support the technology to USB 3.0, and more PCI Express lanes will be curtailed, but the promised support of the same eight-port SATA 6 Gb / sec. It is also reported that neither AMD 1090FX, or AMD 1070 will not support an interface PCI Express 3.0.

Another record for AMD Bulldozer: DDR3 memory overclocked to 3311 MHz

Leading positions Zambezi processors with incredible ease taken in the discipline of CPU-Z, thus wresting the leadership from making Intel. Overclockers opened another discipline in which the Bulldozer can podsidet competitor - «Memory Clock DDR3». The name speaks for itself: to the maximum overclock of DDR3 memory and register the result obtained on the site Hwbot.org. For a long time in this discipline was dominated by the LGA 1156 platform with a processor-based architecture Lynnfield. The maximum result was obtained with the use of frozen storage and Kingston HyperX 1646 MHz was good.

Socket AM3 + platform has registered in the first place «Memory Clock DDR3» overclocker efforts _mat_ , who managed to overclock the memory ADATA Gaming 2.0 Hynix chips to the frequency of 1655.6 MHz (DDR3-3311).

The test stand consisted of a motherboard Asus Crosshair V Formula, and processor FX-8150. Extremely only cooled CPU, memory and fall to a "nitrogen breathing."

Bulldozer processors place prize in the two disciplines of overclocking is very good, but it does not improve their quality of consumer. I'd like to believe that the AMD FX chips will not play the role of a wedding sort of general who invited "on stage" only in exceptional cases.

GPUs with 28nm technology can achieve 45% higher clock frequencies

The rumors about AMD and Nvidia upcoming graphics architectures continue to flow. It has more recently been talk of a near doubling in computing power of AMD's new chips, which according to new data can largely be traced to the GPU frequencies well above 1 GHz.

We have previously taken a closer look at the new 28nm processes such as Taiwan Semiconductor Manufacturing Company (TSMC) will introduce the next six months and what it seems, the semiconductor manufacturer customers very satisfied. TSMC already has more customers for 28nm-production than the company managed to scrape together for several years with their current 40nm technology.

According to information at Web Daily Tech, the new 28-nanometer technology and the fact that TSMC also implemented the High-K Metal Gate (HKMG) in its two more powerful 28nm processes allows us to expect large-dimensioned frequencies.

With TSMC's 28HP technology is expected to AMD and Nvidia to push clock frequencies of its graphics chips by 45% without energy consumption per transistor increases as compared to today's 40nm technology. Could the graphics card manufacturers choose to exploit this surplus to turn up the clock frequencies, we could see the video card with well over 1 GHz GPU frequencies directly from the store shelves.

It is likely that AMD and Nvidia graphics cards in several series also make sure to use the new manufacturing technology, less energy to create more effective graphics solutions. Among other things, we should be able to see even relatively powerful graphics card completely passively cooled, a trend already seen in recent years.

In addition to a huge room for inflated clock speeds, AMD and Nvidia to roll out optimized graphics architectures which include AMD invests in a major refurbishment with GCN architecture (Graphics Core Next). More about this graphics card architecture can be read here .

Unfortunately it will probably be very short on 28nm graphics card in 2011. The best time to launch its first AMD solutions before the new year, but probably we will wait for the major launch until next year, something at least Nvidia is set to.

Processor AMD FX-8150 sets a new record with acceleration up to 8584.8 MHz

It seems, Bulldozer CPU family in general and AMD FX-8150 in particular may soon become a favorite tool, extreme overclockers. Just a few days ago we wrote about the record of this chip overclocked to 8461.51 MHz, effected by the skill of a famous overclocker Andre Young (Andre Yang). However, this well-known enthusiast in all probability, decided not to stop there and made ​​even more impressive result, processor AMD FX-8150 to 8584.8 MHz frequency.

At the same time Andre Young used the same configuration of upper stand as the last time, and for cooling the same way used liquid nitrogen. However, overclockers had to further improve the level of supply voltage on the CPU, which reached, at this time, the turn of 2.076 volts. Largely due to such a high rate of vCore achieved frequency 279.9 MHz bus, which, combined with the value of the multiplier 31X and allowed to achieve the specified outstanding achievement. Recall that in the system in both cases was part of the motherboard ASUS Crosshair V Formula, 2 GB DDR3 modules from Corsair and video card NVIDIA GeForce GTX 580.

AMD FX-8150: Bulldozers enters the market [Test]

TEST Bulldozer is something most tech-savvy heard of. It is AMD's new architecture, built from the ground up, which is the first time since 1999 when they introduced the Athlon series and the K7 architecture, all processors until now been based on. Discussions have been many, so join us in our deep dive of AMD's new performance chip Bulldozer.

When we reviewed AMD Llano, we said that 2011 would be AMD's most important year ever. It is something we stand by today. But in addition to an important year, there has been a very eventful year. AMD chose unexpectedly that their CEO Dirk Meyer, who got the company back on its feet, had to go. Finding a replacement took over six months and the Chosen for the task to bring in AMD in the future, where Rory Read. Read has a background at IBM where he worked 23 years and five years from Lenovo. As COO of Lenovo, he was one of the main people behind the company's success in Asia.

In late September, although Rick Bergman, the company that was their product manager in the consumer market. He was a veteran from ATI, which certainly many old ATI fans knew who it was before AMD bought up the company. Speculation was hot around that he had to leave involuntarily. But it turned out, he left to take a CEO job at Synaptics, which develops touch-products and user interfaces.

Besides having lost two great icons of the company, so today it is launching its third new architecture in less than a year. December 2010 launched their first fusion product: Zacate based on their new Bobcat core and Cedar graphics core. In June came the much talked-about Llano, which is built on AMD's older Stars-cores and Redwood graphics. Today, the company perhaps most anticipated architecture for years, namely the bulldozer that is built on an entirely new processor core. Bulldozers are unlike Zacate and Llano processors without built graphic.

It's certainly not often a product has received so much attention, and AMD has also managed to keep the details secret until the very end. With Bulldozer architecture takes aim at the upper-middle class, and the lower enthusiast segment with what they call the world's first consumer processor. Bulldozers are not only important today, but it is the basic AMD will use additional CPU architectures and APU's. It thus becomes their new K7 who will lead the x86 market underdog into the future.

It is now time to begin the largest plunge in the Bulldozer, the company's most anticipated processor architecture ever.

The information on the chips, AMD Opteron 6000

The network has information about new processors, AMD. It is known that they are designed for servers and will be released later this year.

New items will belong to the generation of Opteron 6000, and their production will be used by architecture Bulldozer.

The published information was eight-Opteron 6212, Opteron 6238, as well as two Opteron 6274 and Opteron 6282 SE.

Will be the most powerful Opteron 6274 and Opteron 6282 SE. Their sixteen cores will be able to operate at a frequency of 2.2GHz and 2.5GHz respectively. Second-level cache will be 8MB (third - 16 MB). In a special mode Turbo Core frequency can reach an impressive 2.8 GHz per core Opteron in 6274, and 3.1 GHz Opteron 6282 SE. If you use one or two cores, these amount to 3.2 GHz frequency and 3.5 GHz respectively.

In 6238 Opteron cache 6MB of second and third level 12 MB, and the base frequency - 2.5 GHz (Turbo Core 3.1GGts in kernel mode, 1.2 3.3 GHz). We Opteron 6212 this figure is 4 MB and 8 MB second third base 2.8 GHz (in the mode to Turbo Core 3.7 GHz).