ToyBrick RK3399Pro Board Shown to Outperform Jetson Nano SBC

Toybrick RK3399Pro

NVIDIA created a lot of buzz when they released $99 Jetson Nano SBC featuring a 128-core Maxwell GPU, and said to deliver 472 GFLOPS of compute performance for running modern AI workloads with a power consumption of around 5 watts. But Jetson Nano is not the only low cost platform to deliver high performance at low power for AI workloads, as for example Rockchip RK3399Pro (RK1808 NPU) found in boards such as Toybrick RK3399Pro is said to deliver 3 TOPS for INT8, 300 GOPS for INT16, and 100 GOPS for FP16 inferences. Those operations per second numbers can be confusing and misleading, so it’s important to check out the performance of actual neural network models, and Rockchip did provide some RK3399Pro benchmarks last year for Inception V3, ResNet34 and VGG16 models comparing the results to Apple A11, Huawei Kirin 970, and NVIDIA Jetson TX2. However, ideally you’d want result from third parties, and Chengwei Zhang got hold of a Toybrick …

Rockchip RK1808 AI Compute Stick Launched with Linux SDK

RK1808 AI compute stick

Rockchip RK1808 looks like a nifty and inexpensive little chip for artificial intelligence applications delivering up to 3.0 TOPS at low power, and the company has already released documentation and a Linux SDK with Caffe and Tensorflow framework support for the chip. So the main hurdle now is to get hardware to play with. Some people are selling (samples?) of the official RK1808-EVB on Taobao, but it costs close to $500 US.  There’s still no RK1808 development board, but Rockchip has discreetly launched the RK1808 AI Compute Stick a few weeks ago. RK1808 AI Compute Stick specifications: SoC – Rockchip RK1808 dual core Cortex-A35 processor with NPU AI inference performance – 3 TOPS for INT8, 300 GOPS for INT16, 100 GOPS for FP16 Video – 1080p60 H.264 decoding, 1080p30 H.264 encoding Camera, ISP – 2MP camera support, ISP with BT.601/656/1120 support Host Interface – USB 3.0 port Power Supply – Via USB port Dimensions – 60 x 19 mm It …

Rockchip RK1808 Datasheet, TRM, Schematics and Linux SDK Released

Rockchip RK1808 Block Diagram

Rockchip RK1808 is the first chip from the company fully dedicated to artificial intelligence applications. The Neural Processing Unit (NPU) features an accelerator delivering up to 3.0 TOPS and is coupled with two low power Arm Cortex-A35 cores allowing it to run Linux. We’ve had the specifications for RK1808 for a while, but the company has recently posted hardware and software resources on their open source website. On the hardware side we’ve got: RK1808 Technical Reference Manual (TRM) RK1808 datasheet PDF Schematics for the company’s official RK1808-EVB On the software side we can get the Linux SDK from Github as explained in the Wiki: Finally configure the build and start the build process: After a while, or more accurately close to two hours on a Laptop with Ryzen 7 2700U processor, 8GB RAM, and hard drive, we’ll get U-boot, Linux, buildroot based rootfs, and firmware files and in IMAGE/RK1808-EVB-V10_20190430.1810_RELEASE_TEST/ directory: Now the only available development board is the RK1808-EVB from …

Rockchip RK1808 Cortex-A35 NPU Delivers up to 3.0 TOPS

RK1808 NPU AIoT Solution

Rockchip RK3399Pro was expected to be the first processor from the company featuring an NPU (Neural Processing Unit) to accelerate artificial intelligence workloads, but eventually went through a redesign, and the community found it gave birth to RK1808 NPU by looking at a device tree file in the Linux kernel. Rockchip has now made the NPU official at CES 2019, and we now know a little bit more. Here are the specifications shared by the company: Dual-core Arm Cortex-A35 CPU NPU computing performance up to 3.0TOPs supporting INT8/INT16/FP16 hybrid operation VPU supporting 1080P video codec Built-in 2MB system-level SRAM Display – MIPI/RGB video output Camera – MIPI/CIF/BT1120 Camera video signal input with built-in ISP Audio Microphone array support with hardware VAD function for low-power monitoring and far-field wake-up Audio output I/O PWM/I2C/SPI /UART USB3.0/USB2.0 PCIe interface Support for Gigabit Ethernet and external WiFi/Bluetooth modules Manufacturing – 22nm FD-SOI process It shares many similar functions as standard media SoCs, but it …

Linux 4.11 Release – Main Changes, ARM & MIPS Architecture

Linus Torvalds has just released Linux 4.11: So after that extra week with an rc8, things were pretty calm, and I’m much happier releasing a final 4.11 now. We still had various smaller fixes the last week, but nothing that made me go “hmm..”. Shortlog appended for people who want to peruse the details, but it’s a mix all over, with about half being drivers (networking dominates, but some sound fixlets too), with the rest being some arch updates, generic networking, and filesystem (nfs[d]) fixes. But it’s all really small, which is what I like to see the last week of the release cycle. And with this, the merge window is obviously open. I already have two pull request for 4.12 in my inbox, I expect that overnight I’ll get a lot more. Linux 4.10 added Virtual GPU support, perf c2c’ tool, improved writeback management, a faster initial WiFi connection (802.11ai), and more. Some notable changes for Linux 4.11 include: …

Linux 4.10 Release – Main Changes, ARM & MIPS Architectures

Linus Torvalds has just released Linux 4.10: So there it is, the final 4.10 release. It’s been quiet since rc8, but we did end up fixing several small issues, so the extra week was all good. On the whole, 4.10 didn’t end up as small as it initially looked. After the huge release that was 4.9, I expected things to be pretty quiet, but it ended up very much a fairly average release by modern kernel standards. So we have about 13,000 commits (not counting merges – that would be another 1200+ commits if you count those). The work is all over, obviously – the shortlog below is just the changes in the last week, since rc8. Go out and verify that it’s all good, and I’ll obviously start pulling stuff for 4.11 on Monday. Linus Linux 4.9 added Greybus staging support, improved security thanks to virtually mapped kernel stacks, and memory protection keys, included various file systems improvements, and …

Rockchip RV1108 Cortex A7 + DSP SoC is Made for Audio & Video Conference and Recording Applications

[Update May 2017: Rockchip has renamed RK1108 to RV1108.] Rockchip has introduced RV1108 ARM Cortex A7 SoC with a 600 MHz DSP targeting visual communication, consumer electronics, automotive DVR, and security applications thanks to its 8-channel I2S audio codec and 1440p H.264 video encoder and decoder. Detailed specifications can be found on the official Rockchip Wiki: CPU – Single-core ARM Cortex-A7 Core processor with NEON and FPU,  32KB/32KB L1 I-Cache/D-Cache, Unified 128KB L2 Cache, and Trustzone Video/Image DSP – Up to 600 MHz, 32KB I-TCM and 32KB I-cache, 128KB D-TCM Memory 12KB internal SRAM DDR3/DDR3L interface – 16 Bits data width, 1 ranks (chip selects), up to 512 MB RAM NAND Flash Interface – 8-bit async NAND flash, 16-bit hardware ECC eMMC Interface – Compatible with standard iNAND interface, eMMC 4.51 standard. SD/MMC Interface – Compatible with SD 3.0, MMC 4.41 System Component 2x 64-bit timers with interrupt-based operation 8x PWMs with interrupt-based operation WatchDog timer Video Video decoder of …