Posted on Monday, July 06 2020 @ 22:21:19 CEST by Thomas De Maesschalck
Kingston has a new gaming keyboard under its HyperX brand, the Alloy Elite 2 Mechanical Gaming Keyboard. This model uses HyperX red linear mechanical switches and has integrated RGB LED lighting. The Alloy Elite 2 is shipping today for $129.99.
HyperX, the gaming division of Kingston Technology Company, Inc., today expanded its Alloy keyboard line up with the release of the HyperX Alloy Elite 2 Mechanical Gaming Keyboard. The new Alloy Elite 2 gaming keyboard is based on the original Alloy Elite design and boasts HyperX red linear mechanical switches built for performance, longevity and an 80 million lifetime click rating per switch. The keyboard includes dedicated media keys with quick access buttons and volume wheel and uses HyperX designed pudding keycaps with translucent dual-layer style and signature HyperX font to enhance LED and RGB light.
The HyperX Alloy Elite 2 features a signature light bar with LED lighting, dynamic RGB lighting effects with per-key lighting and RGB colors. Alloy Elite 2 is built with solid steel frame for durability, offers dedicated media buttons and a large volume wheel along with quick access buttons for brightness, lighting effects, and game mode. Alloy Elite 2 also includes a USB 2.0 pass through, 100 percent anti-ghosting and full N-key rollover functionalities.
“We are excited to announce our new Alloy Elite 2 mechanical keyboard built with HyperX pudding keycaps,” said Jennifer Ishii, keyboard business manager, HyperX. “The Alloy Elite 2 will meet the needs of gamers looking for a fully customizable mechanical gaming keyboard with HyperX switches rated at 80 million clicks per key.”
The Alloy Elite 2 is one of HyperX’s full-featured RGB mechanical gaming keyboards, designed with a durable steel frame for sturdy and stable controls. Users can customize lighting and effects with intuitive NGENUITY software for a brilliant display of colors. The HyperX Alloy Elite 2 also offers onboard memory for keyboard customization allowing users to save up to three profiles. Plus, advanced customization with HyperX NGENUITY software enables users to personalize per-key lighting effects, Game Mode and macro settings. The Alloy Elite 2 is compatible with PC, PS4 and Xbox One.
Posted on Monday, July 06 2020 @ 12:27:53 CEST by Thomas De Maesschalck
There's a lot of hardware out there that never reaches the West. The subject of this article is one of the those things you're unlikely to ever see outside of Asia. Tom's Hardware discovered that Huwaei is shipping a desktop PC that uses the company's eight-core, eight-threaded 7nm Kunpeng 920 ARM v8 processor, in combination with the Huawei D920S10 desktop motherboard, the Yeston RX550 video card, 16GB DDR4-2666 memory, a 256GB SATA HDD, and a 200W PSU. The system runs the 64-bit UOS operating system, which is a Linux distro commissioned by the People's Republic of China to replace foreign operating systems like Windows.
In China, the system sells for 7500 yuan, which is around $1,060. It's quite slow compared to typical x86 CPUs though, and only seems good for light office work and media consumption:
The system did run a Blender BMW test render, but it completed in 11 minutes and 47 seconds, which is woefully slower than most modern chips. The system did well streaming 4K video, but 'choked' during local playback due to poor encoding performance. The narrator says the system is obviously best for light office work only.
Posted on Monday, July 06 2020 @ 11:59:59 CEST by Thomas De Maesschalck
An entry of the Intel Core i7-1165G7 processor in the Geekbench database reveals this Tiger Lake chip packs quite a punch. TechPowerUp goes as far as to say that the 1165G7 mauls AMD's Ryzen 7 4700U "Renoir" APU in most tests. The 1165G7 is a quad-core, eight-threaded 10nm part so it's pretty interesting to see it beating AMD's eight-core, eight-threaded 4700U in both single-threaded and multi-threaded performance:
A Geekbench comparison between two Lenovo laptops, one powered by an i7-1165G7, and the other by a 4700U, shows a staggering 36.8% performance lead for the Intel chip in single-threaded performance, while also being 0.5% faster in multi-threaded performance. The i7-1165G7 features a 4-core/8-thread CPU with "Willow Cove" cores, while the 4700U lacks SMT, and is an 8-core/8-thread chip with "Zen 2" CPU cores. The game changes with the Ryzen 7 4800U, where the 8-core/16-thread chip ends up 22.3% faster than the Core i7-1165G7 in the multi-threaded test owing to SMT, while Intel's single-threaded performance lead is lowered to 29.3%.
We're looking forward to in-depth reviews of Tiger Lake!
Posted on Monday, July 06 2020 @ 11:50:53 CEST by Thomas De Maesschalck
Samsung researcher published a paper about amorphous boron nitride (a-BN), a new material that holds promise for use in future generations of semiconductors. The South Korean tech giant claims the material is ideal for widespread application in future DRAM and NAND solutions:
Researchers at the Samsung Advanced Institute of Technology (SAIT) have unveiled the discovery of a new material, called amorphous boron nitride (a-BN), in collaboration with Ulsan National Institute of Science and Technology (UNIST) and the University of Cambridge. Published in the journal Nature, the study has the potential to accelerate the advent of the next generation of semiconductors.
2D Materials – The Key to Overcoming Scalability Challenges
Recently, SAIT has been working on the research and development of two-dimensional (2D) materials – crystalline materials with a single layer of atoms. Specifically, the institute has been working on the research and development of graphene, and has achieved groundbreaking research outcomes in this area such as the development of a new graphene transistor as well as a novel method of producing large-area, single-crystal wafer-scale graphene. In addition to researching and developing graphene, SAIT has been working to accelerate the material’s commercialization.
“To enhance the compatibility of graphene with silicon-based semiconductor processes, wafer-scale graphene growth on semiconductor substrates should be implemented at a temperature lower than 400°C.” said Hyeon-Jin Shin, a graphene project leader and Principal Researcher at SAIT. “We are also continuously working to expand the applications of graphene beyond semiconductors.”
2D Material Transformed – Amorphous Boron Nitride
The newly discovered material, called amorphous boron nitride (a-BN), consists of boron and nitrogen atoms with an amorphous molecule structure. While amorphous boron nitride is derived from white graphene, which includes boron and nitrogen atoms arranged in a hexagonal structure, the molecular structure of a-BN in fact makes it uniquely distinctive from white graphene.
Amorphous boron nitride has a best-in-class ultra-low dielectric constant of 1.78 with strong electrical and mechanical properties, and can be used as an interconnect isolation material to minimize electrical interference. It was also demonstrated that the material can be grown on a wafer scale at a low temperature of just 400°C. Thus, amorphous boron nitride is expected to be widely applied to semiconductors such as DRAM and NAND solutions, and especially in next generation memory solutions for large-scale servers.
“Recently, interest in 2D materials and the new materials derived from them has been increasing. However, there are still many challenges in applying the materials to existing semiconductor processes.” said Seongjun Park, Vice President and Head of Inorganic Material Lab, SAIT. “We will continue to develop new materials to lead the semiconductor paradigm shift.”
Posted on Monday, July 06 2020 @ 11:41:44 CEST by Thomas De Maesschalck
Facebook published a paper (PDF) that describes a machine learning-based approach to upsample images. The neural supersampling from Facebook seems similar to NVIDIA's Deep Learning Super Sampling, but it does not require any proprietary hardware or software. Facebook will reveal the technique at SIGGRAPH 2020.
Closest to our work, Nvidia has recently released deep-learned supersampling (DLSS) [Edelsten et al. 2019] that upsamples low-resolution rendered content with a neural network in real-time.
In this paper, we introduce a method that is easy to integrate with modern game engines, requires no special hardware (e.g., eye tracking) or software (e.g., proprietary drivers for DLSS), making it applicable to a wider variety of existing software platforms, acceleration hardware and displays.
We observed that, for neural supersampling, the additional auxiliary information provided by motion vectors proved particularly impactful. The motion vectors define geometric correspondences between pixels in sequential frames. In other words, each motion vector points to a subpixel location where a surface point visible in one frame could have appeared in the previous frame. These values are normally estimated by computer vision methods for photographic images, but such optical flow estimation algorithms are prone to errors. In contrast, the rendering engine can produce dense motion vectors directly, thereby giving a reliable, rich input for neural supersampling applied to rendered content.
Our method is built upon the above observations, and combines the additional auxiliary information with a novel spatio-temporal neural network design that is aimed at maximizing the image and video quality while delivering real-time performance.
At inference time, our neural network takes as input the rendering attributes (color, depth map and dense motion vectors per frame) of both current and multiple previous frames, rendered at a low resolution. The output of the network is a high-resolution color image corresponding to the current frame. The network is trained with supervised learning. At training time, a reference image that is rendered at the high resolution with anti-aliasing methods, paired with each low-resolution input frame, is provided as the target image for training optimization.
Target applications for Facebook may include AR and VR applications for its Oculus platform.
Via: a href="https://wccftech.com/neural-supersampling-is-a-hardware-agnostic-dlss-alternative-by-facebook/" target="_blank">WCCF Tech
Posted on Monday, July 06 2020 @ 11:30:35 CEST by Thomas De Maesschalck
VideoCardz reports about a listing of an AMD Zen 3-based processor. Codenamed "100-000000114-07_22/15_N", this appears to be an engineering sample of the "Milan" EPYC processor.
The codename of the sample suggests it has a 1.5GHz base clock and a 2.2GHz Boost. That seems really low but it's important to remember that engineering samples of the previous-gen EPYC processors showed similar clocks, while final versions had much higher frequencies. Milan is a 7nm+ part and it should deliver an uptick in performance/Watt.
AMD Milan is also known as Genesis (internal codename), is the upcoming Zen3 EPYC series (likely to be called 7xx3), expected in late 2020. AMD has confirmed that the Milan-based EPYC series will ship “late 2020”.
Here we go ????
AMD Eng Sample: 100-000000114-07_22/15_N
1.5GHz base, 2.2GHz boost. This sample seems to match the 64-core Rome ES' in terms of clockspeed
Posted on Monday, July 06 2020 @ 10:51:06 CEST by Thomas De Maesschalck
NVIDIA fan CaddDood uploaded a beautiful render of the NVIDIA Geforce RTX 3080 Ti on the NVIDIA subreddit. This is a fan-made render of what the card may look like, based on earlier leaks that emerged last month.
The first Ampere-based consumer video cards from NVIDIA are expected later this year. We don't know an exact launch date but it's believed that the green team will make an announcement in September. It's believed that Ampere will deliver a big gain in real-time ray-tracing performance.
The card was modeled in SolidWorks, imported to Cinema 4D and rendered with Redshift 3.0.23 in 8K resolution. A couple more screenshots of CaddDood's mockup can be viewed at Imgurg.
Posted on Monday, July 06 2020 @ 10:33:19 CEST by Thomas De Maesschalck
A Geekbench entry reveals Intel is working on the Core i9-10850K. This is a ten-core, twenty-threaded processor with an unlocked multiplier, a 3.6GHz base clock, and 5.2GHz Turbo Boost Max 3.0 frequency. The nominal clockspeed is 100MHz below the 10900K and it's unknown if it has a Thermal Velocity Boost (TVB).
TechPowerUp speculates the chip may lack TVB altogether. The site believes the 10850K may be an OEM variant with lower cooling requirements than the 10900K.
It wouldn't surprise us if the processor lacked TVB altogether. It's likely that the i9-10850K is an OEM-exclusive targeted at pre-built designers that don't want to deal with the steep cooling requirements of the i9-10900K to give end-users visible boosting to its TVB Max frequencies of 5.30 GHz. The i9-10850K offers nearly identical Geekbench performance to the i9-10900K.
Posted on Monday, July 06 2020 @ 10:28:58 CEST by Thomas De Maesschalck
While AMD motherboards have historically been cheaper than Intel counterparts, this is now not longer the case if you need a Micro-ATX or Mini-ITX motherboard. ExtremeTech analyzed the motherboard market and concludes the price gap between Intel and AMD motherboards gets larger the smaller you go.
For ATX form factor motherboards, the price difference is mostly a wash. If you need a Micro-ATX motherboard, the cheapest AMD option carries a $20 premium over Intel. Those looking to build compact Mini-ITX systems can expect even bigger price differences.
The site notes that the Asrock A320M-ITX should be the cheapest Mini-ITX option for the AMD platform, with a MSRP of $99. Unfortunately, this motherboard is nowhere to be found. The cheapest Mini-ITX board you can actually buy right now is the Asrock X570 Phantom Gaming ITX/TB3, at $240. That's roughly $140 more than the cheapest Mini-ITX options for the Intel platform.
So why are there no low-cost small form factor motherboards for AMD's Ryzen? ExtremeTech speculates it could be due to strong demand for these boards, which may be emptying retail store shelves quickly. Alternatively, it could also be due to the coronavirus pandemic:
There are basically two reasons why AMD boards might be running expensive. First, it’s possible that some coronavirus-related inventory issues have resulted in better Intel board availability than AMD availability. We only see real gaps in mATX and mini-ITX. If manufacturers have cut back on building boards that AMD buyers don’t choose all that often, it would explain the limited inventory and high prices.
It will be interesting to see how this situation evolves over the coming months and years.
Posted on Monday, July 06 2020 @ 10:16:38 CEST by Thomas De Maesschalck
While malware for Apple's Mac platform is rare, it's important to remember that Macs aren't magically immune to attacks. Four years ago the first ransomware for Mac was discovered and now security researchers discovered "ThiefQuest", a piece of malware that basically consists of a spyware and a ransomware component.
Patrick Wardle, principal security researcher at the Mac management firm Jamf, speculates someone was making a piece of spyware with the ability to completely remote control an infected Mac, and then later decided to add a ransomware component as a way to make some money.
Overall, the risk seems low as ThiefQuest is primarily spread via pirated software. Security researchers note the primary mains of distribution seems to be via torrents. To get infected, you need to download pirated software and run the infected installer. Then you also need to dismiss a string of security warnings from macOS.
Though ThiefQuest is packed with menacing features, it's unlikely to infect your Mac anytime soon unless you download pirated, unvetted software. Thomas Reed, director of Mac and mobile platforms at the security firm Malwarebytes, found that ThiefQuest is being distributed on torrent sites bundled with name-brand software, like the security application Little Snitch, DJ software Mixed In Key, and music production platform Ableton. K7's Devadoss notes that the malware itself is designed to look like a "Google Software Update program." So far, though, the researchers say that it doesn't seem to have a significant number of downloads, and no one has paid a ransom to the bitcoin address the attackers provide.