Microprocessor Report (MPR) Subscribe

Current MPR Articles

October 27, 2014

  • Soft Machines Targets IPC Bottleneck (MPR)
    A new technology called VISC promises greater performance by breaking down a single thread and executing it on multiple cores. The technology works in hardware on any program and any instruction set.
  • AppliedMicro ARMs for Embedded (MPR/NWR)
    AppliedMicro’s new Helix family of ARMv8-compatible embedded processors is based on the company’s X-Gene server chips and also inherits DNA from its older PacketPro communications chips.

October 20, 2014

  • ARC HS38 Can Run High-Level OS (MPR)
    The new HS38 CPU core from Synopsys can run full versions of embedded Linux using its MMU, cache-coherent SMP, optional L2 cache, and other new features.

October 13, 2014

October 6, 2014

  • Cortex-M7 Doubles Up on DSP (MPR/MCR)
    ARM's new 32-bit Cortex-M7 aims to bring higher-performance digital signal processing to MCUs for high-end embedded applications. The new architecture delivers 47% better compute performance than Cortex-M4.

September 29, 2014

  • MIPS Warrior Joins 64-Bit Battle (MPR/MCR)
    The I6400 CPU is the first licensable core in Imagination Technologies’ Warrior family to implement a 64-bit architecture. As an interAptiv core, it fits in the middle of the MIPS lineup.

September 22, 2014

  • Sparc64 XIfx Uses Memory Cubes (MPR)
    Fujitsu’s new Sparc64 XIfx supercomputer processor uses Micron’s Hybrid Memory Cubes and numerous other improvements to push system-level performance toward 100 petaflops.

September 15, 2014

  • Broadwell-Y Launches 14nm Era (MPR)
    Core M targets 2-in-1 PCs and tablets, fitting into a thermal envelope of just 4.5W TDP while offering a preview of mainstream 14nm Broadwell products.
  • Intel’s 14nm Graphics Broadside (MPR/MCR)
    Broadwell delivers a next-generation GPU with 20% better compute performance, 50% more texturing, and OpenCL 2.0 compatibility. This GPU will scale from high-end PCs to low-cost tablets.

September 8, 2014

September 1, 2014

  • X-Gene 2 Aims Above Microservers (MPR)
    AppliedMicro's 28nm X-Gene 2, which is currently sampling, delivers Xeon E3–class compute performance, Xeon E5–class memory bandwidth, and highly integrated, low latency 10GbE networking.
  • Intel Buys Axxia for Base Stations (MPR/NWR)
    Intel is aiming to make a bigger splash in wireless infrastructure by acquiring the former LSI networking business. It faces many execution challenges that could hinder ongoing success, however. 
  • Samsung First With 20nm Processor (MPR/MCR)
    The Exynos 5430 and 5433 are the first application processors in 20nm, providing power and performance advantages over 28nm products. Samsung also announced its first LTE-Advanced modem chip.

August 25, 2014

August 18, 2014

  • Nvidia’s First CPU Is a Winner (MPR/MCR)
    Nvidia's 64-bit Denver CPU uses dynamic instruction translation to outperform other mobile processors while offering full ARMv8 compatibility. It will appear in Tegra K1-64 later this year.
  • The Case for Open Instruction Sets (MPR)
    Two UC Berkeley professors make the case that an open instruction-set architecture would enable free competition in CPU design and that RISC-V is the best open ISA available.
  • The Case for Licensed Instruction Sets (MPR)
    ARM makes the counterargument that a licensed instruction set fills a commercial need for support, reliability, and compatibility, enabling a robust ecosystem that provides value for an ISA. 

August 11, 2014

  • Broadcom Boosts Small Cells (MPR/NWR)
    Broadcom's new BCM617x5 wireless base-station processors add crest-factor reduction, better carrier aggregation, more-powerful CPUs and DSPs, and support for China Mobile’s Zuc.
  • Marvell Puts IoT on a Chip (MPR)
    Marvell has combined a basic ARM microcontroller with Zigbee, Bluetooth, and Wi-Fi radios to create a family of processors for the Internet of Things.

August 4, 2014

Purchase an Article