Linley Newsletter: January 24, 2019

 weSRCH's Best of the Internet Award

Linley Newsletter

Please feel free to forward this to your colleagues

Issue #636

January 24, 2019

Independent Analysis of Microprocessors and the Semiconductor Industry

Editor: Tom R. Halfhill

Contributors: Linley Gwennap, Mike Demler, Bob Wheeler

In This Issue:

- Analysts' Choice Winners for 2018

- Intel Turns Graphics up to (Gen) 11

- Microchip Previews RISC-V FPGAs

Last Chance for Discount on New Deep Learning Report

We have completely revised our report "A Guide to Processors for Deep Learning" to reflect the growth in this rapidly evolving market. AI has been dominating the news with dozens of new products and technologies announced in just the last year. Deep learning isn't limited to just data centers and self-driving cars, it is even moving into client devices such as smartphones and embedded (IoT) systems.

This new edition focuses on the hardware technologies and products driving this market. You'll find detailed technical coverage of announced chip products from AMD, Eta Compute, Graphcore, Gyrfalcon, Habana, Intel (including former Altera, Mobileye, Movidius, and Nervana technologies), Mythic, NXP, Nvidia (including Tegra and Tesla), Qualcomm, Wave Computing, and Xilinx. It also covers IP cores from AImotive, Arm, Cadence, Cambricon, Ceva, Imagination, Synopsys, Videantis, and the open-source NVDLA. Other chapters cover Google's TPU family of ASICs and Microsoft's Brainwave.

Hurry! Order by January 31 to take advantage of the introductory offer. For more information, visit our website:

Analysts' Choice Winners for 2018

By The Linley Group

To recognize the top semiconductor offerings, The Linley Group presents its 2018 Analysts' Choice Awards. This year's winner for Best Embedded Processor is Intel's Myriad X. It adds video and crypto accelerators along with the Neural Compute Engine, which boosts AI performance to one trillion operations per second (TOPS). The design consumes a mere 2.5W (typical), suiting it for embedded, IoT, and mobile systems.

The open-source NVDLA wins our Best Processor IP Award. The NVDLA stands out for its technical features, which include configurable hardware for all neural-network layers, a choice of integer and floating-point data types, tensor-processing functions, and weight-compression algorithms.

We chose Qualcomm's Snapdragon 710 as Best Mobile Processor. It was the industry's first product employing two big CPUs and six little ones. At its release, it boasted a significant lead over other mid-premium designs in graphics performance, LTE data rates, and Wi-Fi data rates.

We picked the Zynq UltraScale+ RFSoC as Best Networking Chip. This unique design represents a new class of FPGAs that boost integration by including high-performance analog-to-digital converters (ADCs). In communications, the Xilinx chip serves in remote radio heads for wireless infrastructure as well as in cable modems for wired broadband.

For Best Server Processor, we chose Nvidia's Tesla T4. Based on the Turing architecture, the T4 features 320 tensor cores that can each multiply two 4x8 matrices in a single cycle. The accelerator card delivers about 3x better AI performance per watt than its predecessor, the Tesla P4, and about 20x better than the Skylake-SP processors from Intel.

TSMC wins our Best Technology Award for its 7nm FinFET process. The company not only beat other foundries to this node but also for the first time surpassed Intel, once the unchallenged industry leader, putting TSMC more than a year ahead of all competitors.

Microprocessor Report subscribers can access the full article:

Intel Turns Graphics up to (Gen) 11

By David Kanter

The upcoming 10nm Ice Lake processors will employ Intel's first mainstream GPU with 1Tflop/s of single-precision compute capability. The Gen11 graphics architecture boosts power efficiency through clever features and fixed-function hardware for the latest media formats. The new GPU will be Intel's first to reach production since Gen9 debuted in the 2015 Skylake processors. (The company developed a Gen10 GPU for Cannon Lake, but 10nm-process delays sank that product.) As a result, AMD's integrated Radeon GPU outmatches Intel's current integrated GPUs for mainstream processors.

Intel has tweaked each graphics generation for greater scalability. Gen11's large performance boost owes to a repartitioned hierarchy of shader cores and memories. The company has also accelerated its adoption of new architectural features. For example, Skylake's Gen9 GPU was first to employ full conservative rasterization, beating AMD and Nvidia to market. Although Intel invented coarse pixel shading, the 10nm delays will prevent Gen11 from being the industry's first architecture with this feature. Nevertheless, the new GPU adds this capability along with tile-based rendering.

The company is continuing to emphasize media processing by adding fixed-function support for new formats such as VP9 and several HDR varieties, along with greater throughput and power efficiency. On the display side, it boosted the pixel output to drive larger displays and adopted variable refresh rates to more intelligently operate displays in conjunction with the GPU.

Because of its PC market share, Intel's mainstream graphics define the industry's performance floor. Gen11 is scheduled to arrive in Ice Lake in 4Q19, delivering more than 1Tflop/s -- a 10x increase from the first on-die GPUs in 2011 -- as well as a tremendous improvement in programmability and other features. Ice Lake should also reach graphics parity with AMD, and higher-performance versions will once again nip at the low end of the discrete-graphics market.

Microprocessor Report subscribers can access the full article:

Microchip Previews RISC-V FPGAs

By Bob Wheeler

Moving from RISC-V evangelist to committed customer, Microchip is developing SoC FPGAs using cores from SiFive. With most versions of its PolarFire FPGA now shipping, the former Microsemi is previewing what's next while stopping short of a product announcement. Due to sample in 2H19, the SoC FPGAs combine the PolarFire FPGA design with SiFive's U54-MC complex, comprising quad 64-bit application cores, a 64-bit monitor core, and a 2MB L2 cache.

The PolarFire SoC architecture includes a hardened 32-bit DDR4/LPDDR4 controller as well as processor-boot memories. This design enables the U54-MC processor complex to boot and run without consuming any FPGA resources. Although Microchip withheld details, the processor can also directly access certain I/Os; PolarFire includes, for example, dual PCI Express Gen2 controllers.

Although Microchip is first to integrate hard RISC-V CPUs, SoC FPGAs are far from new. Since 2013, the company (then Microsemi) has shipped SmartFusion2 FPGAs with an integrated Arm Cortex-M3 microcontroller. Sporting up to 150,000 LUTs, SmartFusion2 sits below PolarFire in logic density. Meanwhile, both Intel (Altera) and Xilinx have shipped multiple SoC-FPGA generations that integrate Arm Cortex-A CPUs capable of running Linux.

Microprocessor Report subscribers can access the full article:

About Linley Newsletter

Linley Newsletter is a free electronic newsletter that reports and analyzes advances in microprocessors, networking chips, and mobile-communications chips. It is published by The Linley Group. To subscribe, please visit:

Domain: Electronics
Category: Semiconductors

Recent Newsletters

Linley Newsletter: August 8, 2019

Linley Newsletter Please feel free to forward this to your colleagues Issue #664 August 8, 2019 Independent Analysis of Microprocessors and the Semiconductor Industry E

08 August, 2019

Linley Newsletter: August 1, 2019

Linley Newsletter Please feel free to forward this to your colleagues Issue #663 August 1, 2019 Independent Analysis of Microprocessors and the Semiconductor Industry E

01 August, 2019

Linley Newsletter: July 25, 2019

Linley Newsletter Please feel free to forward this to your colleagues Issue #662 July 25, 2019 Independent Analysis of Microprocessors and the Semiconductor Industry

25 July, 2019