Microprocessor Report (MPR) Subscribe

Current MPR Articles

January 23, 2017

  • i.MX8 Masters Multimedia
    NXP’s new i.MX8/8M processors combine up to three different types of ARM cores and various multimedia blocks to address consumer, industrial, and automotive applications. 
  • NXP Upgrades Radar Processor
    The S32R27 quadruples the performance of NXP’s earlier automotive radar processors using a Power-based architecture in a microcontroller configuration with embedded memory.
  • Ambiq Rides Quadratic Power Curve
    Ambiq has developed an ultra-low-power microcontroller using subthreshold-voltage circuits, extending battery life without sacrificing performance.
  • AMD Hones Learning Instincts
    Leaping into the hot machine-learning market, AMD unveiled the Radeon Instinct GPU line, topped by the MI25 training accelerator and the MI6 and MI8 inference accelerators.

January 16, 2017

  • Analysts’ Choice Winners for 2016
    The Linley Group recognizes 2016’s top products in the categories of embedded, mobile, networking, server, processor IP, and related technology. And the winners are…

January 9, 2017

  • Year in Review: Clouds Open Up for New Processors
    Although 2016 was a quiet year for server processors, vendor activity sowed the seeds of change. ARM processors began generating revenue and customers increasingly employed designs that use accelerators.
  • Snapdragon 835 First to 10nm
    Qualcomm is the first vendor to offer a 10nm smartphone processor, extending battery life. Unlike its predecessor, the 835 uses fairly standard Cortex CPUs and integrates a Wi-Fi baseband.
  • Samsung 10nm Takes Density Crown
    Samsung is the first company to reach 10nm production, offering the densest FinFET process to date. The 10nm LPE process will power smartphone and server processors in 2017. 
  • GlobalFoundries Offers 56Gbps Serdes
    A year after launching its 14nm FX-14 FinFET process for ASICs, GlobalFoundries is offering serdes intellectual-property cores that support both 56Gbps PAM4 and 28Gbps NRZ coding.

January 2, 2017

  • Broadcom Builds KBP in 16nm
    Broadcom is shipping the industry’s first search coprocessor built in FinFET technology. The KBP (knowledge-based processor) search engines appear in service-provider routers and switches as an adjunct to OEMs’ own ASICs.
  • Helio Rises to X27
    The X23 and X27 add new deca-core models to the Helio smartphone-processor lineup. The X27 boosts the Cortex-A72 CPUs to 2.6GHz, while the X23 provides a lower-price option.

December 26, 2016

  • Draft 802.11ax Up for 2017 Ballot
    The IEEE’s 802.11 group has published draft 1.0 of 802.11ax, a new 2.4/5GHz standard that increases Wi-Fi speeds while also improving spectral efficiency.
  • Editorial: More MPR for You
    We are expanding the number of stories available to our subscribers to better cover the broader set of technologies that affect microprocessor design.
  • Qualcomm Samples Centriq for Servers
    Centriq is the first server processor to be developed in 10nm technology, putting the 48-core design ahead of other ARM-based competitors as well as Intel on this metric.

December 19, 2016

  • Macom to Grab AppliedMicro
    High-speed-analog specialist Macom has offered $776 million for AppliedMicro, a developer of communications and processor products. The company will immediately divest AppliedMicro’s X-Gene server processors.
  • Cortus APS29 Doubles Instruction Fun
    Cortus has delivered its APS29 intellectual-property (IP) core, which supports dual instruction issue and provides greater single-core performance than the older APS25.

December 12, 2016

  • Goodix Shifts to Fingerprint Sensors
    The small Chinese company has become a big hit, shipping in the past year more than 120 million fingerprint sensors with innovative features such as Live Finger Detection.
  • FD.io Software Vectorizes Packets
    Fast Data I/O (FD.io), a Linux Foundation project that launched in 1Q16, champions parallelization of packet processing. It’s based on a 15-year-old Cisco technology called vector packet processing (VPP).
  • Many Options for Machine Learning
    From Siri to self-driving cars, machine learning and neural networks are rapidly growing in popularity, and engineers are using everything from CPUs to specialized hardware to keep pace.

December 5, 2016

  • Broadcom Broadens StrataDNX
    Broadcom is closing out 2016 with a significant expansion of its StrataDNX line, including two new line-card devices and two new Qumran switch chips.
  • Everspin MRAM Targets Enterprise
    Everspin’s new non-volatile memory offers DRAM-like performance as well as greater endurance and retention than flash memory, targeting embedded and enterprise storage.
  • Samsung Gear S3 Uses Exynos 7270
    A tiny new Exynos chip for wearables uses 14nm technology to combine a simple application processor with an LTE modem. A suite of new connectivity chips complements the design.

November 28, 2016

  • Mellanox Debuts 200Gbps InfiniBand
    Mellanox has retaken the lead in low-latency-interconnect speeds by introducing its Quantum switch and ConnectX-6 adapter for the High Data Rate (HDR) 200Gbps InfiniBand (IB) standard.
  • Editorial: Nanometer Madness
    The foundries no longer follow traditional metrics in labeling their process nodes, making IC-manufacturing progress more difficult to accurately assess.
  • Huawei First With Cortex-A73, Bifrost
    HiSilicon’s Kirin 960, shipping in phones soon, delivers leading-edge features including ARM’s newest CPU and GPU designs, 4x LTE carrier aggregation, and dual-camera support.
  • Mali-G51 Brings Bifrost to Mainstream
    The new Mali-G51 is ARM’s smallest GPU to support the Khronos Vulkan API. The new Mali-V61 video engine scales from 1080p HD to 4K UltraHD resolution.

November 21, 2016

  • New Standards Open Server Platform
    Non-Intel servers will benefit as OpenCAPI and CCIX compete to become the standard coherent interconnect while Gen-Z aims to simplify deployment of innovative memory and storage technologies.
  • GC8000 GPU Shoots for the Stars
    VeriSilicon has delivered the GC8000 GPU, an evolutionary update offering better performance per unit area and higher throughput.
  • DeepBench Measures Deep Learning
    Baidu has released a collection of over 150 neural-network kernels designed to evaluate the performance of various processors, along with performance data for Intel and Nvidia hardware.
  • Aquantia Pushes Copper to 100Gbps
    QuantumStream promises 100Gbps per lane over existing direct-attach cabling (DAC) for 3m server-to-ToR links. Aquantia is targeting customer production in 2018.

November 14, 2016

  • Qualcomm-NXP: A New Goliath
    Qualcomm’s proposed $47 billion acquisition of NXP will strengthen its mobile offering while diversifying the company, giving it strong positions in automotive, embedded, and IoT.
  • Cortex-M23/M33 Bolster IoT Security
    ARM has rolled out the first Cortex-M CPUs to implement the ARMv8-M ISA, as well as a new CoreLink system interconnect, a security-IP core, and a complete subsystem for IoT processors.
  • Marvell Aims for Enterprise 25GbE
    Marvell’s new Prestera CX 84xx Bobcat 3 switches and Alaska C 88X51xx PHYs target private-cloud data centers upgrading from 10GbE to 25GbE, a market the company says its competitors have neglected.
  • Splashdown for Intel’s Apollo Lake
    After a big graphics boost, a new Atom processor supports low-cost PCs under the Pentium and Celeron brands as well as embedded markets using the Atom E3900 and A3900 names.

November 7, 2016

  • Apple Turbocharges PowerVR GPU
    According to our analysis, Apple has customized Imagination’s GPU to achieve very high clock speeds, delivering a 50% performance gain on some tests but falling short on others.
  • Adreno 530 Tops Mobile GPUs
    In a comparison of flagship smartphone processors, Qualcomm’s custom GPU design leads ARM’s Mali and Imagination’s PowerVR in both performance and die size.
  • ST’s Mostly Modern Microcontroller
    STMicroelectronics’ new STM32H7-series extends the performance of the company’s high-end microcontroller line by moving to a new 40nm embedded-flash process.
  • Quantenna IPO Raises Profile, Cash
    Quantenna raised $107 million in an October 28 IPO that hit the top of its anticipated price range. The company is also first to announce chips for the draft 802.11ax standard.

October 31, 2016

  • Tomahawk II Doubles 100GbE Ports
    Broadcom is aiming to reinforce its dominance in top-of-rack Ethernet switching by sampling the 16nm BCM56970 Tomahawk II, the first switch to offer 64x100GbE ports.
  • Snapdragon 427 Moves to Category 7
    Qualcomm’s three new smartphone processors (Snapdragon 427, 626, and 653) provide minor improvements, refreshing the company’s lineup in time for the holiday buying season.
  • Adjusting SPEC CPU2006 Scores
    Advanced compilers inflate SPEC benchmark scores by 8–23% using optimizations that are irrelevant to most server users. Adjusting these scores enables more accurate comparisons of server processors.
  • Ceva X1 Controls Cellular IoT Devices
    Ceva’s new X1 is a synthesizable processor core for cellular-connected IoT and wearable devices that use Narrowband IoT (NB-IoT) or other low-power communications protocols.
  • MIPS I6500 Juices Bandwidth
    Imagination has released the I6500 CPU IP, updating its midrange MIPS design to improve bandwidth and multicluster coherence.

More Articles

Purchase an Article


 

Events

Linley Cloud Hardware Conference 2017
Covers system design for cloud computing and networking
February 8, 2017
Hyatt Regency Hotel, Santa Clara, CA
Register Now!
More Events »

Newsletter

Linley Newsletter
Analysis of new developments in microprocessors and other semiconductor products
Subscribe to our Newsletter »