Intel releases a new generation of Xeon D processors
On February 25, Intel released a new Xeon D processor, including the Xeon D-1700 and Xeon D-2700 series, to replace the Xeon D-1600 and Xeon D-2100 series. Intel says the new products feature industrial-grade reliability and a variety of hardware-based security features to meet the computing needs of space- and power-constrained environments.
The new Xeon D processor continues the Ice Lake architecture, known as Ice Lake-D, replacing the previous Broadwell architecture with significantly improved performance and configuration. It uses the Sunny Cove microarchitecture, manufactured on a 10nm process, supports three/four channels of DDR4-3200 memory, provides up to 56 PCIe channels, and supports 100Gbe network connectivity. The Xeon D-1700 series offers 4- to 10-core products, while the Xeon D-2700 series is 4-core to 20-core, with a total of 36 models. The highest end of these is the Xeon D-2799, which has 20 cores and 40 threads, the base frequency is 2.4GHz, the turbo frequency is 3.4GHz, equipped with 30MB of L3 cache, and the TDP is 129W.
Intel said that the new Xeon D processor has a 2.4x increase in visual processing inference performance, up to 1.7x higher for complex network workloads such as 5G UPF at the edge, up to 1.5x for SD-WAN, SASE, and edge use cases using IPSec, up to 1.8x for application delivery controllers and TLS security appliance use cases, and a 1.56x improvement for communication use cases.
As a fork of the Xeon family of processors, Xeon D is primarily aimed at the "microserver" market and is suitable for dense computing environments where power and space are limited but require higher performance. Positioned between the Atom and Xeon E3 series, it competes with Arm-architected server processors.
Redmi K50 Series image specification exposure
On February 25, the xiaomiui.net exposed the image parameters of the Redmi K50 series.
According to xiaomiui.net, the Redmi K50 series codenames are matisse, rubens, munch, and the corresponding models are Redmi K50, Redmi K50 Pro and Redmi K50 Pro+.
Among them, redmi K50 is equipped with Qualcomm Snapdragon 870 flagship processor, rear 48 million main camera, 8 million ultra-wide angle and macro triple camera, Redmi K50 Pro equipped with Tianji 8100 processor, rear 48 million main camera, 8 million ultra-wide angle, the other camera details are not exposed.
Redmi K50 Pro+ is equipped with a Tianji 9000 processor, a rear 108 million pixel main camera, model Samsung HM2, and the details of the other two cameras are not exposed.
From the specification point of view, the K50 Pro+ is equipped with 108 million pixels is the highest quality Sensor among the three models, its unit pixel area is 0.7μm, the sensor size is 1/1.52 inch, it supports dual-core focus and 9 pixels in 1, and the equivalent unit pixel area still reaches 2.1μm, which can take photos of extreme low light at 12 million pixels.
In addition, the Redmi K50 series adopts E4 AMOLED flexible direct screen, which supports up to 120W super second charge, and the machine is expected to be released in March.
Armored Man introduces a new generation of UFS flash memory
The flash memory of the mobile phone is also crucial for the overall performance and fluency, and the current UFS 3.1 flash memory continuous read and write speed has been more than GB/s, which has a fight with the desktop SSD. Now, Jiaoxia has announced the launch of a new generation of UFS flash memory, the industry's first support for MIPI M-PHY v5.0, performance increase of up to 90%, capacity up to 512GB.
The U.S. subsidiary today announced that it will sample the industry's first MIPI M-PHY v5.0-compliant UFS flash memory samples, using the company's BiCS Flash 3D flash technology, providing 128GB, 256GB and 512GB capacities.
The new generation of UFS flash memory provides high-speed read and write performance, mainly for a variety of mobile devices, including smartphones.
In HS-GEAR5 mode, the theoretical interface speed is as high as 23.2Gbps per channel, and 46.4Gpbs under x2 channels - for reference, PCIe 3.0 hard disks provide interface speeds up to 32Gbps, and PCIe 4.0 x4 is 64Gbps.
Armor said that compared to the previous generation of devices, the sequential read and write performance of 256GB devices is about 90% and 70%, respectively. In addition, the random read and write performance of the 256GB device is approximately 35% and 60% better than the previous generation of devices, respectively.
According to The Armored Man, samples of the 256GB version of the flash memory will be shipped on February 25, and other products will start to follow up in August, and it is not surprising that there will be flagship phones using this flash memory as soon as the second half of this year.
The AMD Ryzen™ 7000 Series will integrate a GPU
As more and more news about the Ryzen™ 7000 series processors on the market, more and more heavy news has also appeared in the public eye.
According to the revelations KOMACHI_ENSAKA, AMD's next-generation Ryzen 7000 "Raphael" CPU will integrate a new RDNA 2 graphics core architecture, and it is no longer limited to APU.
It is reported that the upcoming Ryzen 7000 series processor allows the GPU to integrate 4 RDNA 2 computing units, for reference, the Ryzen 6000 series processor has a maximum of 12 RDNA 2 computing units.
In addition, the god of revelations also said that the AMD Raphael series does not have high performance requirements for GPUs, and its launch purpose is more for users to use as a basic graphics processing tool, but the follow-up does not rule out that AMD has an upgrade plan.
MediaTek Tianji 8100 exposed
MediaTek Tianji 9000 has been mass-produced for commercial use, and the first model is OPPO Find X5 Pro Tianji Edition. After the Tianji 9000, MediaTek Tianji 8000 series chips are about to debut, and it is benchmarked against the Qualcomm Snapdragon 888.
Today, some digital bloggers have given the parameters of mediatek Tianji 8100 chip, which uses TSMC's 5nm process, consisting of 4 Cortex A78 large cores, 2.85 GHz CPU frequency and 4 Cortex A55 small cores, CPU main frequency of 2.0GHz, GPU G610 MC6.
In terms of running points, the comprehensive score of An Tutu of Tianji 8100 exceeded 820,000 points, surpassing the Snapdragon 888.
In addition, the news pointed out that the Redmi K50 Pro is equipped with the Tianji 8100 chip, and the super cup K40 Pro+ released by Redmi last year is equipped with the Snapdragon 888, which means that the performance of the Redmi K50 Pro surpasses the Redmi K40 Pro+.
The opportunity was released in March, it is worth noting that the Redmi K50 series to be released this time in addition to the K50 Pro, there may be K50 and K50 Pro+, etc., of which the K50 is equipped with Snapdragon 870, and the K50 Pro+ may be equipped with Tianji 9000.
Intel 15th Generation Core Roadmap Leaked
Apple's self-developed M-series processors have had a major impact on Intel, and they are also trying to break this gap and try to reverse the unfavorable effect.
Judging by the newly exposed roadmap, Intel hopes to compete with Apple's 14-inch MacBook Pro through its Arrow Lake series (loaded with M1 Pro or M1 Max chips), while the 15th-generation Arrow Lake processors may be delivered in late 2023 or early 2024, prioritizing minimum energy consumption to deliver high performance.
The roadmap also shows that Intel will leverage TSMC's 3nm process process. Apple currently uses the 5nm process in its latest chips, and it is expected to adopt 3nm in 2023.
At present, in the book data, Intel has surpassed the "M1 Max" chip. Benchmarks show that its latest Core i9 processor scored higher on tests than Apple's M1 Max chip, but it still lags behind in terms of power consumption ratio, with a 4 percent performance boost offset by a significant reduction in battery life.
Tests have shown that laptops equipped with the latest i9 chips have only 6 hours of video playback time. In contrast, Apple officially labeled the latest 16-inch MacBook Pro to offer up to 21 hours of battery life (playing offline videos).