Intel Neural Compute Stick 2 with Myriad X VPU Finally Announced

Intel Neural Compute Stick 2

Intel Neural Compute Stick was first introduced in early 2017 as a USB compute that allows AI inference at the edge with low power consumption. The stick is based on Myriad Movidius 2 VPU (Vision Processing Unit), and was found to significantly improve inference performance on Raspberry Pi 3 board over a proprietary GPU accelerated solution. However, a little later last year, Intel also announced Movidius Myriad X VPU with claims of up to 10 times DNN performance over Myriad 2 VPU. But so far, we would only see solutions launched with the latter, and it looks like Intel is finally ready to bring Myriad X VPU to the market with the company announcing Intel Neural Compute Stick 2 at Intel’s artificial intelligence (AI) developer conference in Beijing taking place on November 14 and 15. Intel NCS 2 (Neural Compute Stick 2) specifications: Processor – Intel Movidius Myriad X Vision Processing […]

Samsung Exynos 9820 Application Processor Supports 2 Gbps LTE, 8K Videos, UFS 3.0 Storage

Exynos 9820

Samsung has just unveiled a new high-end Exynos processor that comes with an enhanced NPU (Neural Processing Unit), and an LTE Advanced Pro modem supporting 2.0 Gbps downloads. Exynos 9820 is a tri-cluster octa-core processor with two custom Arm cores, two Cortex-A75 cores, and four Cortex-A55 cores coupled with an Arm Mali-G76 MP12, and manufactured using Samsung’s latest 8nm LPP FinFET process, which reduces power consumption by up to 10 percent compared to 10nm LPP process. Samsung Exynos 9820 key features and specifications: CPU – Dual-core custom CPU +  dual-core Cortex-A75 + quad-core Cortex-A55 GPU – ARM Mali-G76 MP12 Integrated NPU Memory I/F – LPDDR4x Storage I/F – UFS 3.0, UFS 2.1 Display – WQUXGA (3840×2400), 4K UHD (4096×2160) Cellular Connectivity LTE Modem – LTE-Advanced Pro Cat.20 8CA 2Gbps (DL) / Cat.20 3CA 316Mbps (UL) 4×4 Multiple-Input, Multiple-Output (MIMO), 256-QAM (Quadrature Amplitude Modulation), and Enhanced Licensed-Assisted Access (eLAA) technology. Camera […]

96Boards AI Sophon Edge Developer Board Features Bitmain BM1880 ASIC SoC

96boards Sophon Edge

Bitmain, a company specializing in cryptocurrency, blockchain, and artificial intelligence (AI) application, has just joined Linaro, and announced the first 96Boards AI platform featuring an ASIC: Sophon BM1880 Edge Development Board, often just referred to as “Sophon Edge”. The board conforms to the 96Boards CE specification, and include two Arm Cortex-A53 cores, a Bitmain Sophon Edge TPU delivering 1 TOPS performance on 8-bit integer operations, USB 3.0 and gigabit Ethernet. Sophon Edge specifications: SoC ASIC – Sophon BM1880 dual core Cortex-A53 processor @ 1.5 GHz, single core RISC-V processor @ 1 GHz, 2MB on-chip RAM, and a TPU (Tensor Processing Unit) that can provide 1TOPS for INT8,and up to 2 TOPs by enabling Winograd convolution acceleration System Memory – 1GB LPDDR4 @ 3200Mhz Storage – 8GB eMMC flash + micro SD card slot Video Processing – H.264 decoder, MJPEG encoder/decoder, 1x 1080p @ 60fps or 2x 1080p @ 30fps H.264 decoder, […]

Rockchip RK3399Pro EVB and NPU Performance Demos (Video)

Rockchip RK3399Pro EVB

Rockchip RK3399Pro was announced as an updated version of RK3399 processor with an NPU (Neural Processing Unit) capable of delivering 2.4 TOPS for faster A.I. workloads such as face or object recognition. There haf been some delays in the past because of a redesign of the processor that placed the NPU’s RAM on the PCB instead of on-chip for cost reasons. Eventually we got more details about RK3399Pro, and today I also received a 15-page presentation with some more information about the software, and processor itself. But even more interesting, that’s the first time I see Rockchip’s official RK3399Pro EVB (Evaluation Board), and the guys at Khadas uploaded a video to explain a bit more about the board, and showcase the NPU performance measured up to 3.0 TOPS with an object recognition demo, and an “body feature” detection demo – for the lack of a better word – running in […]

Khadas Edge RK3399/RK3399Pro Board Crowdfunding Campaign Launched

Khadas Edge is another upcoming Rockchip RK3399 board that we initially covered this summer, and comes with a particular design in the sense it is both a single board computer with USB and Ethernet port, and a system-on-module through its edge connector. Later on we found out, the company was working on a variant called Khadas Edge-V with the edge connector being replaced by a more standard 40-pin header, as well as Khadas Edge1S powered by Rockchip RK3399Pro for AI applications. The company has now launched the three boards and accessories through an Indiegogo crowdfunding campaign. I’ve already covered the two RK3399 boards in previous post, and  the RK3399Pro based Khadas Edge-1S (2GB RAM, 2x MIPI-CSI, …) will be similar to Edge board but with , so instead I’ll have a closer look at the specifications for Khadas Captain carrier board for Khadas Edge / Edge-1S: Edge connector – 314-pin […]

MediaTek Helio P70 Launched as an Upgrade to Helio P60 Processor

MediaTek Helio P70

MediaTek Helio P60 processor was introduced earlier this year as one of the first Arm Cortex A73 processors from the company, and also integrated NeuroPilot AI technology for faster or more complex A.I. workloads. The company has now launched an upgrade with Helio P70 that comes with many of the same features combined with incremental performance and power consumption improvements. MediaTek Helio P70 key features and specifications with highlights in bold showing differences over Helio P60: CPU – big.LITTLE octa-core with four Arm Cortex-A73 up 2.1 GHz and four Arm Cortex-A53 up to 2.0 GHz GPU – Arm Mali-G72 MP3 at 900MHz Multi-core AI processing unit (Mobile APU) – 280 GMAC/s; offers deep learning facial recognition, object and scene identification, user behavior-informed performance and other AI and AR application enhancements. System Memory – Up to 8GB, dual-channel LPDDR4x @ 1800 MHz, up to 4GB LPDDR3 @ 933 MHz Storage –  […]

Qualcomm Snapdragon 675 Cortex-A76 based Processor Targets High-Tier Mid-Range Smartphones

Snapdragon 675

Last August, Qualcomm unveiled Snapdragon 670 processor with performance and efficiency Kryo 360 cores believed to be based on Cortex-A75 and Cortex-A55 cores, but the company has just announced an update with Snapdragon 675 mobile platform powered by Kryo 460 cores which are based on Arm Cortex-A76  according to Anandtech, and so far, Cortex-A76 was only found in “premium” processors such as HiSilicon Kirin 980. Beside the upgrade cores, the company also claims the processor brings “outstanding gaming, a leap in artificial intelligence (AI) capability and a cutting-edge camera”. Snapdragon 675 main specifications: CPU 2x Kryo 460 (Cortex-A76 based) @ 2.0GHz with 256KB L2 6x Kryo 460 (Cortex-A55 based) @ 1.8GHz with 64KB L2 GPU – Adreno 612 DSP – Hexagon 685, Qualcomm AI Engine Display – Support for FHD+ display Camera – Spectra 250L ISP with triple-camera support Audio – Qualcomm aptX and Aqstic Modem – Snapdragon X12 LTE […]

More Details about Rockchip RK3399Pro SoC, and RK1808 NPU

RK3399Pro Block Diagram

First announced in January 2018, Rockchip RK3399Pro was supposed to be a pin-to-pin compatible with Rockchip RK3399 processor, and adding a Neural-Network Processing Unit (NPU) capable of delivering 2.4 TOPS for acceleration A.I. workloads. Shortly after Pine64 announced they’d be offering Rockpro64-AI board in August, and later on Vamrs unveils ROCK960 PRO at a Linaro Connect event with an expected Q2 2018 launch. But none of the RK3399Pro boards are available, as there have been delays with RK3399Pro, and some commented an external NPU would be launched first with further details.   But today – courtesy of Vamrs – we have some more details about RK3399Pro features, a likely explanation for the delay, and some information about Rockchip RK1808 NPU chip. Contrary to the CES 2018 announcement, Rockchip RK3399Pro will come in a 27x27mm FCBGA1372 package instead of the 22x22mm FCBGA828 package for RK3399. So pin-to-pin compatibility is out of […]

Exit mobile version
UP 7000 x86 SBC