CNX Software - Embedded Systems News

April 29, 2021April 29, 2021 by Jean-Luc Aufranc (CNXSoft) - 9 Comments

Allwinner V831 NPU (Neural Processor Unit) reverse-engineered

When Sipeed introduced MAIX-II Dock AIoT vision development kit, they asked help from the community to help reverse-engineer Allwinner V831‘s NPU in order to make an open-source AI toolchain based on NCNN. Sipeed already had decoded the NPU registers, and Jasbir offered help for the next step and received a free sample board to try it out. Good progress has been made and it’s now possible to detect objects like a boat using cifar10 object recognition sample. Allwinner V831’s NPU is based on a customized implementation of NVIDIA Deep Learning Accelerator (NVDLA) open-source architecture, something that Allwinner (through Sipeed) asked us to remove from the initial announcement, and after reverse-engineering work, Jasbir determined the following key finding: The NPU clock defaults to 400 MHz, but can be set between 100 and 1200 MHz NPU is implemented with nv_small configuration (NV Small Model), and relies on shared system memory for all […]

April 29, 2021April 30, 2021 by Jean-Luc Aufranc (CNXSoft) - 25 Comments

Allwinner Processor 2021-2022 Roadmap – Allwinner T827, T723 and T1033 SoC’s

Last year, we published the Allwinner A-series processor roadmap for 2020-2021 with notably Allwinner A33E, A100, and A200 SoCs. Allwinner A100 was supposed to be released in 2019, but a search on Aliexpress showed exactly zero matches. We did write about Allwinner A100 mainline Linux support a little while ago, and today, I eventually found one Allwinner A100 tablet with the $60 Hyundai HyTab 7WC1. I had no better luck in my search for Allwinner A33E and A200 platforms. But I also got lucky today with another Allwinner roadmap for 2021-2022 with some very interesting new processors, provided they happen. A look at 2020 processors But let’s look at the year 2020 first. We already have most details about Allwinner A133 quad-core Cortex-A53 processor, which is also called T509 apparently following the merging of the Allwinner A-Series business unit, focusing on tablets, into the Allwinner T-Series for industrial & automotive […]

April 29, 2021 by Jean-Luc Aufranc (CNXSoft) - 1 Comment

MPi4 NEC MediaPlayer Kit slot-in card powers 4K digital signage displays with Raspberry Pi CM4

NEC Display Solutions has been integrating Raspberry Pi Compute Modules into commercial displays used for digital signage and presentation platforms for over 5 years now. The company, now called Sharp NEC Display Solutions, has launched a new Intel Smart Display Module (SDM)-like slot-in card based on Raspberry Pi Compute Module 4 called “MPi4 NEC MediaPlayer Kit” and designed as an entry-level card alternative to the company’s Intel SDM slot-in cards designed for NEC large format 4K displays. MPi4 NEC MediaPlayer Kit specifications: SoM – Raspberry Pi Compute Module 4 “NEC Edition” with SoC – Broadcom BCM2711 quad-core Arm Cortex 72 @ 1.5GHz, VideoCore IV GPU with OpenGL ES 3.0 graphics, H.265 (HEVC) up to 4Kp60 decode, H.264 up to 1080p60 decode, 1080p30 encode System Memory – 4GB LPDDR2 SDRAM Storage – 32 GB eMMC flash Dimensions – 55 × 40 × 4.7 mm Interface Board for Raspberry Pi Compute Module […]

April 28, 2021April 28, 2021 by Jean-Luc Aufranc (CNXSoft) - 2 Comments

Xilinx Introduces Kria K26 SoM and vision AI devkit based on Zynq Ultrascale+ XCK26 FPGA MPSoC

Silicon vendors will usually focus on chip design, and provide an expensive evaluation kit to early customers, leaving the design of cost-optimized boards and system-on-modules to embedded systems companies. But Xilinx has decided to enter the latter market with the Kria portfolio of adaptive system-on-modules (SOMs) and production-ready small form factor embedded boards starting with Kria K26 SoM powered by Zynq UltraScale+ XCK26 FPGA MPSoC with a quad-core Arm Cortex-A53 processor, up to 250 thousand logic cells, and a H.264/265 video codec designed for Edge AI applications, as well as computer vision development kit. Kria K26 System-on-Module Kria K26 module specifications: MPSoC – Xilinx Zynq Ultrascale+ custom-built XCK26 with quad-core Arm Cortex-A53 processor up to 1.5GHz, dual-core Arm Cortex-R5F real-time processor up to 600MHz, Mali-400 MP2 GPU up to 667MHz, 4Kp60 VPU, 26.6Mb On-Chip SRAM, 256K logic cells, 1,248 DSP slices System Memory – 4GB 64-bit DDR4 (non-ECC) Storage – […]

April 28, 2021 by Jean-Luc Aufranc (CNXSoft) - 2 Comments

19-inch rackmounts support up to 16 front-removable Raspberry Pi SBC’s

We previously covered Myelectronics’ 19-inch rackmount taking up to 12 Raspberry Pi boards. The solution is nice to keep a bunch of Raspberry Pi 4 in a small space and expose the HDMI ports as full HDMI ports. But what happens if you need to replace a Raspberry Pi board, or simply change the MicroSD card? Yes, you’d typically need to turn off all boards, remove the rack mount, replace the board/install the MicroSD card, and reconnect everything. That’s probably why the company has now launched two new 19-inch 2U and 3U rack mounts respectively supporting up to 16 front-removable Raspberry Pi (model B) single board computers. That works a bit with a typical NAS, where you insert or remove a hard drive by pulling it out and back. The 2U rackmount pictured above comes with front-removable brackets where you can install the Raspberry Pi, as well as mounting screws, […]

April 28, 2021 by Jean-Luc Aufranc (CNXSoft) - 14 Comments

Tool-free industrial RJ45 connector speeds up cable assembly by up to 25%

You’d normally need a crimp tool to make or repair your own RJ45 Ethernet cables. The process is split into two parts with cable preparation for straight or crossed cables, and then assembly with the crimp tool, before testing your cable. Harting RJ Industrial Multifeature is an RJ45 connector that is described as “tool-free”, and “robust and reliable: with integrated cutting tools for fast and easy assembly. So the crimp tool is not needed anymore. The company explains it typically takes 50% of the time for the preparation of the cable and another 50% for connecting it to a traditional RJ45 plug. But the RJ Industrial Multifeature is said to save around 25% of the assembly time, by making the second part faster. You’d still need to prepare the cable in the traditional manner, but then the second part of the assembly should be quite faster and less prone to […]

April 27, 2021 by Jean-Luc Aufranc (CNXSoft) - 4 Comments

Jasper Lake Mini PC offered with up to Celeron N5095 processor

Intel officially launched the Jasper Lake processors at the beginning of the year. The new family provides an upgrade to the Gemini Lake processors that have been found in low-cost low power mini PC in the last few years. While there have been some leaks of the ASUS PN41 Jasper Lake mini PC a couple of months ago, no manufacturer has officially introduced this type of computer so far. But Newsmay Technology has now started to promote its AC8 mini PC with a choice of the six 6W/10W Celeron/Pentium processors previously announced, as well as a 15W Celeron N5095 processor that I had never heard of. AC8 mini PC specifications: Jasper Lake SoC (one or the other) Intel Celeron N4500 dual-core processor @ 1.1/2.8GHz with UHD graphics; 6W TDP Intel Celeron N5100 quad-core processor @ 1.1/2.8GHz with UHD graphics; 6W TDP Intel Pentium Silver N6000 quad-core processor @ 1.1/3.1GHz with […]

April 27, 2021April 27, 2021 by Jean-Luc Aufranc (CNXSoft) - 7 Comments

SiFive Intelligence X280 64-bit RISC-V processor integrates AI extensions

The last RISC-V core announced by SiFive was the U8-Series out-of-order RISC-V Core IP that aims to compete against Arm Cortex-A72 Core. But in their latest announcement, the company built upon the 64-bit RISC-V U7-series with the SiFive Intelligence X280 multi-core, Linux capable RISC-V processor adding vector extensions and SiFive Intelligence Extensions, and optimized for AI/ML compute at the edge. SiFive Intelligence X280 key features: 64-bit RISC-V ISA with 8-stage dual-issue in-order pipeline, coherent multi-core, Linux capable based on U7 series core. SiFive Intelligence Extensions for ML workloads – BF16/FP16/FP32/FP64, int8 to 64 fixed-point data types 512-bit vector register length – Variable-length operations, up to 512-bits of data per cycle High-performance vector memory subsystem Memory parallelism provides cache miss tolerance Virtual memory support with precise exceptions Up to 48-bit addressing SiFive Intelligence includes software solutions to leverage the X280’s features and provide “great AI inference performance” using TensorFlow Lite. No […]