A new piece of information about AMD Van Gogh ultra-low power 6-15 W APUs has appeared on the Web. These chips will find use in ultrabooks or other compact technology and should be presented by the end of 2021.
The Van Gogh family will use two key AMD technologies: Zen 2 for the processor and RDNA 2 for graphics. It is believed that the design will be different from all previous APUs, but retain the monolithic construction of the Renoir or Cezanne. The Van Gogh chips will have a BGA FF3 design and will be manufactured using a 7nm process technology.
Van Gogh APUs have four AMD Zen 2 cores. In general, the die area should be as small as possible, which, among other things, will be achieved by reducing L2 and L3 caches. No information has been received regarding dual-core models. But AMD is striving to make the most of the rejected chips to expand the product line downward, so their exit is not excluded.
Unofficial AMD Processor Roadmap
There is no data yet regarding the RDNA 2 graphics unit. The number of computational modules for senior models will not exceed 8-10 CU (up to 512-640 stream processors). Despite the new architecture, you should not expect miracles in iGPU performance, as the low TPP will not allow you to reach high frequencies.
A month ago, there was a rumor about a 256-bit (or four-channel) DDR5 memory controller from Van Gogh. According to Moore’s Law is Dead, this is not true, because this decision will lead to an increase in the cost and complexity of the chip, while AMD, on the contrary, is moving towards simplification and reduction of the area. Van Gogh is much more likely to rely on LPDDR5 memory.
Chinese firm Netac has confirmed the research and development of DDR5 memory that will be able to operate at an effective frequency of 10 GHz. For the first DDR5 batches, this figure will be at least 4800 MHz (this value is mentioned in the specifications of Intel Alder Lake processors), but the roadmaps list the range up to 8400 MHz. As for DDR5-10000, such modes are extreme and require at least manual overclocking or a tuned XMP profile.
Netac has been on the RAM market recently – since 2018. To date, their portfolio includes not only mainstream RAM products, but also gaming sets, including DDR4 with RGB backlighting. As for the new standard, Netac has already received the first batches of the Micron DDR5 chip labeled IFA45 Z9ZSB. Initially, the modules will operate in DDR5-4800 mode with CL40-40-40 delays, according to the JEDEC standard.
As we know, the release of the first Intel LGA1700 platform and CPU Alder Lake with DDR5 support will take place in the second half of this year. AMD should release a response in the form of Socket AM5 and Zen 4 processors, but that will likely only happen next year.
We expected that AMD’s Zen 2 powered Ryzen chips in the Surface Laptop 4 would be superior to the Surface Laptop 3. But so much? […]
Microsoft unveiled the Surface Laptop 4 on Tuesday with some bold performance claims: The Surface Laptop 4 is said to outperform the Surface Laptop 3 by around 70 percent, and in certain scenarios even offer twice the performance.
Does it live up to the hype? Does it ever do that? Although the difference between the two generations of Surface Laptops becomes smaller in everyday activities, our early tests show that the Surface Laptop 4 wins all CPU-intensive tasks by a wide margin. (If you’ve been keeping an eye on Microsoft’s offer of a free pair of Surface Earbuds for every Surface Laptop 4 preordered by April 15th, this could be an incentive).
Dispatched April 27th
Microsoft will begin shipping the Surface Laptop 4 in Europe on April 27th in two traditional display sizes: 13.5-inch and 15-inch. Inside you have the choice between two microprocessors: either the 11th generation core chip (Tiger Lake) from Intel or the Surface Edition processor, which was jointly developed by Microsoft and AMD, and on AMD’s mobile Ryzen 4000 CPUs (Zen-2 architecture) is based. While the Intel CPUs are off the shelf, the Surface Edition processors are highly clocked parts that cannot be found in any other notebook in the world.
We received a 15-inch version of the Surface Laptop 4 in which the eight-core Ryzen 7 4980U Surface Edition is installed. The operating system reports it as a 2 GHz chip, with boost speeds that (according to the CPU-Z utility) don’t seem to go higher than 4 GHz.
This article is just a glimpse of how the Surface Laptop 4 performed on a limited number of benchmarks before our full review appears. You may notice slight differences from the numbers we ultimately publish in our review as we typically run multiple instances of each benchmark to ensure accuracy. We haven’t run any battery life tests either.
We’re also artificially narrowing the scope of this performance preview to show how the Surface Laptop 4 compares to the various iterations of the Surface Laptop 3 and other Microsoft devices. Our next review of the Surface Laptop 4 will include competing third-party notebooks to show how it compares to a broader field.
Surface Laptop 4: Early Performance Benchmarks
UL’s PCMark 10 measures everything from how your laptop is performing in a video chat, to word processing and spreadsheets, photo editing and rendering, to light gaming. There is hardly a generation difference in these everyday tasks: The Surface Laptop is 4 to 6 percent ahead of the fastest Surface Laptop 3.
Since CPU performance is so critical to everyday laptop performance, we’ll test it in two ways. Maxon’s Cinebench offers a simple, comparative metric for evaluating laptops, desktops and individual processors. Here the Surface Laptop 4 offers a little more than double (104 percent) the CPU performance of the fastest Surface Laptop 3.
HandBrake, an open source utility, represents our other CPU test by putting an extended load on the laptop – dubbing a full-length, high-definition Hollywood movie into a format that is safe to store on an Android tablet suitable is. This process can take more than an hour with older hardware. How well can the Surface Laptop 4 handle full power over a long period of time? How well did Microsoft engineers design the Surface Laptop 4’s cooling solutions so that it doesn’t throttle to maintain stability? HandBrake tells us: very good!
With HandBrake, the Surface Laptop 4 does the job in about half the time (49 percent) that the Surface Laptop 3 takes.
In our performance test, we use UL’s other benchmark, 3DMark, to assess graphics performance. The Surface Edition processor offers integrated graphics instead of a discrete GPU. The Surface Laptop 4 shows a more modest increase in graphics performance of 33 percent compared to the similarly equipped Surface Laptop 3. (The benchmark crashed when running on the Surface Laptop 1 and Surface Laptop 2, which is why the results are not shown for these models).
As the picture (and the Surface Pro 7+ score) suggests, the Surface Laptop 4 with the strongest graphics will likely be the Intel Core-based version with the built-in Xe GPU. We hope to be able to test this version of the Surface Laptop 4 shortly.
Conclusion: A strong start for the Surface Laptop 4
These preliminary benchmarks give some very positive indications of the latest Microsoft Ryzen 7 4980U Surface Edition processor powering the Surface Laptop 4. Many questions remain unanswered, especially the all-important battery life. We also didn’t look closely at whether the Surface Laptop 4 throttled its CPU or GPU under different loads. We haven’t received an Intel-based Surface Laptop 4 to test either – how well will that work? Last but not least, we of course also have to evaluate the display quality, keyboard and trackpad performance, connectivity and other properties of the Surface Laptop 4.
We’ll answer all of these questions and more in our official review. In the meantime, we have this strong indicator: twice as fast as the Surface Laptop 3. Twice as fast. In most benchmarks, Microsoft’s Ryzen-powered Surface Laptop 4 does what it says it is.
* As Managing Editor of PCWorld, Mark Hachman’s focus is on Microsoft news and chip technology, among other things.
Microsoft Corporation and Nuance Communications Inc. announced a definitive agreement under which Microsoft will acquire Nuance for $ 19.7 billion. The Redmond-based tech giant will pay $ 56 per share, 23% more than its market value at its close at the end of last week. The deal promises to be one of the largest in Microsoft history.
Nuance Communication is a leading US technology company and a pioneer in conversational AI and ambient intelligence technologies. In particular, Nuance develops voice recognition systems, and one of the most famous projects is the Dragon family of software products.
The software uses deep learning algorithms to decode speech and over time improves accuracy by adapting to the user’s voice. For example, Apple’s personal assistant Siri is based on Dragon and other Nuance designs. According to the press release, 77% of US hospitals and 85% of Fortune 100 companies around the world trust Nuance’s range of services.
Microsoft’s acquisition of Nuance builds on a successful partnership between the companies that began in 2019. By complementing the Microsoft Cloud for Healthcare with Nuance solutions, and through Nuance’s expertise and relationships with systems vendors Electronic Health Records, Microsoft will expand the e-health marketplace, and the total Microsoft and Nuance targeting market in this area will double to nearly $ 500 billion.
“Nuance brings AI to healthcare and is a pioneer in the practical application of enterprise AI,” said Satya Nadella, CEO of Microsoft. “Artificial intelligence is the most important technology priority, and healthcare is the most urgent application. Together with our partner ecosystem, we will place advanced AI solutions in the hands of professionals around the world to drive better decision making and create more meaningful connections as we accelerate the growth of Microsoft Cloud for Healthcare and Nuance. ”
The deal was unanimously approved by the boards of directors of Nuance and Microsoft. It is projected to close by the end of the calendar year and is subject to Nuance shareholder approval, certain regulatory approvals and other normal closure conditions.
Consolidated revenue of contract manufacturer Taiwan Semiconductor Manufacturing Company for March and the first quarter of 2021 reached a record high, according to Taiwanese publication DigiTimes. For example, the turnover in March increased by 21.2% compared to February and by 13.7% year on year to 129.13 billion Taiwan dollars (4.54 billion US dollars).
According to experts, TSMC’s revenue for March 2021 exceeded the previous record set in September 2020. At that time, the contract manufacturer reached a peak in the production of Huawei orders, but then tougher US sanctions against the Chinese company came into force.
TSMC’s total revenue increased 16.7% YTD to 362.41 billion Taiwan dollars ($ 12.74 billion) from a year earlier, also the highest quarterly.
In the coming days, TSMC management will hold a profit conference to discuss the results of its operations in the first quarter and present the prospects for business development. Financial records are expected to be broken several times this year as demand for semiconductor products is huge, coupled with rising prices and the removal of discounts.
The German firm Alphacool has unveiled four large radiators for uncompromising water cooling. All major parts of the new products are made of copper, including heat exchange chambers, channels and fins. Only mounting plates and elements of the radiator frame are made of steel, threaded connections are made of brass.
Radiators NexXxoS Monsta 200 mm (price from 85 euros) and Monsta 400 mm (from € 116) are designed for fans with a standard size of 180 mm or 200 mm and have a thickness of 86 mm. Both products have seven standard G1 / 4 holes. Six of them are used to connect hoses, and the seventh port is located in the end chamber and can only be used for filling, draining or venting the system.
The density of the fins is 12 FPI, which allows the use of fans with lower static pressure and lower speed.
NexXxoS UT60 Nova 1080 mm (price from 138 euros) is one of the most efficient radiators in CBO. The huge cooling surface is comparable to three standard 360mm radiators and requires nine 120mm fans per kit.
This makes it easy to dissipate more than 1000 watts of excess heat using extremely quiet or silent Carlsons. For example, this amount of heat is equivalent to one 64-core AMD Threadripper 3000 and at least two GeForce RTX 3090 graphics cards.
NexXxoS XT45 Quad 480/560 mm (price from 90 euros) is a junior version of the UT60 Nova 1080 mm. It comes with two mounting plates, each of which can accommodate four 120mm or 140mm fans.
The novelty corresponds to the cooling capacity of two 240/280 mm radiators. NexXxoS XT45 Quad 480/560 mm is capable of meeting the needs of any desktop system with one processor and one graphics card.
Alphacool also introduced Apex Liquid ECO, which is “environmentally friendly and easy to dispose of”. Purified water is the basis, and additional additives have an antibacterial effect and reduce the surface tension of the coolant, which facilitates the elimination of bubbles in the circuit.
Internet providers have been facing delays in the supply of new routers for more than a year due to a global shortage of chips, according to Bloomberg. According to people familiar with the issue, the lead time for orders reaches 60 weeks, which is more than double the previous values.
According to Karsten Geveke, head of the European regional division of Zyxel Communications Corp, the abrupt cessation of production a year ago due to the coronavirus was exacerbated by an unprecedented surge in demand for broadband equipment and other electronics due to the transfer of employees to remote work, and students and students to distance learning.
Since January, Zyxel has been asking customers to order products a year in advance because lead times for components such as Broadcom chips have doubled. However, there is another side to the coin: over the past twelve months, Zyxel shares are up more than 181%.
US networking equipment maker Adtran has also warned customers about supply chain risks and extended delivery times in recent months. A company spokesman said the company has expanded its UK warehouse space, more than doubling its inventory and logistics capacity to avoid problems. The container ship Ever Given also played a role in the supply crisis, blocking the Suez Canal for a week.
By the way, the situation with routers can directly affect the Ukrainian market. Networked devices have lower margins than smartphones or computers. In turn, products for less affluent markets such as Eastern Europe use simple electronic components with even lower margins. Obviously, less profitable work is relegated to the background. Therefore, in this situation, small telecom operators or consumers in the countries of the second or third world will suffer the most.
ASUS continues to release themed merchandise for fans of the Gundam anime universe to celebrate its 40th anniversary. The new product will be the ROG Strix XG279Q-G Gundam Edition gaming monitor. Most likely, it will be possible to buy a monitor on free sale only in Japan or a number of other Asian countries.
ROG Strix XG279Q-G Gundam Edition is 27-inch and features Fast IPS panel with 170Hz refresh rate and 1ms response time. The monitor is classified as G-Sync Compatible, which means it is compatible with VESA Adaptive-Sync and AMD FreeSync Premium.
The matrix has a resolution of WQHD (2560 x 1440 pixels), a response time of 1 ms (gray to gray), a contrast ratio of 1000: 1 and a standard brightness of 400 cd / m², which corresponds to the DisplayHDR 400 standard. The coverage of the sRGB color space is 130%, and DCI-P3 – 95%.
The monitor is equipped with Extreme Low Motion Blur, GameFast Input, Shadow Boost, Low Blue Light, Dynamically Local Dimming, Flicker-free and others. External interfaces include one DisplayPort 1.2, two HDMI 1.4, USB hub, headphone output and a pair of 2W stereo speakers.
The stand can be adjusted for height, tilt and rotation to portrait mode. Gundam styling and RGB lighting can be seen on the back and leg. There was no information on the recommended price.
At the disposal of the web edition VideoCardz was a confidential slide from Intel with the main characteristics of the future server platform Eagle Stream. Together with it, the chipmaker will release Xeon Scalable processors from the Sapphire Rapids family. These CPUs will mark the transition to DDR5 RAM in the server segment and will bring with them a number of important innovations.
Intel Sapphire Rapids engineering sample in LGA4677 design
Intel Sapphire Rapids processors will offer up to 56 physical cores and are available in four- and eight-socket configurations. The equipment of the “stones” includes an 8-channel DDR5-4800 controller, up to 64 GB of HBM2E buffer memory with a bandwidth of 1 TB / s and up to 80 PCI Express 5.0 lines. The nominal thermal pack for the flagship solutions will be 350 watts.
Other features of Sapphire Rapids include support for new instruction sets, the Compute Express Link 1.1 interface and the 10nm Enhanced SuperFin process technology. It should be added that Intel intends to use these processors both in the server (Sapphire Rapids-SP) and in the HEDT segment (Sapphire Rapids-X). Nothing has been heard about the latter yet, but according to VideoCardz informants, work on motherboards for them has already begun.
Last October, AMD officially announced the purchase of FPGA manufacturer Xilinx for the equivalent of $ 35 billion. Although in this case we are not talking about real money, but about a significant part of AMD shares (~ 25%), which will be transferred to the holders of Xilinx securities. The combined corporation is expected to have a market capitalization of $ 135 billion and a combined market of over $ 110 billion.
Today it became known that the shareholders of both companies voted to approve the merger of Xilinx and AMD. In fact, this is one of the important stages on a long journey towards the completion of the financial and organizational process. The closing of the deal continues to be subject to many other conditions being met, including obtaining the necessary approvals from US regulators. The management of the companies continues to count on the final merger by the end of 2021, although a delay until 2022 is possible.
“Over the years, AMD has successfully pursued a long-term growth strategy and deepened partnerships to ensure leadership in high performance computing,” said Lisa Su, President and CEO of AMD. “The acquisition of Xilinx marks the next step on our path to making AMD the preferred strategic partner for the world’s largest and most important technology companies as an industry leader with the vision, talent and scale to support their future innovations.”