Posts

Showing posts from March, 2019

Utilizing the Raspberry Pi GPU Horsepower

Image
Doing a little research on special purpose computation, I found it is possible to reprogram the Raspberry Pi GPU to do some powerful calculations. The Raspberry Pi 12 cores GPU (more properly called Quad Processing Units (QPUs) ) it has theoretical max performance of 24 GPFLOS    Theres an advanced book titled:    Raspberry Pi GPU Audio Video Programming by Jan Newmarch    in which he states: " The GPU is capable of 1 Gpixel/s, 1.5 Gtexel/s, or 24 GFLOPs of general-purpose compute and features a bunch of texture filtering and DMA infrastructure "    Here's a list of related articles and code libraries: Broadcom released documentation on their proprietary GPU some time ago, the file is no longer available on their site, some of the material below may be based on that dated material.    GPU Accelerated Object Recognition on Raspberry Pi 3 & Raspberry Pi Zero (not source so far)       Hacking The GPU F...

Human Activity Detection Methods

Image
 Monitor human activity for security, profit or senior health monitoring  Microsoft Knect   Initially bundled with the Xbox360 and later made available for PC for more utilitarian purposes.   Wikipedia A Virtual Dressing Mirror was a common commercial Kinect This demo was posted by a Chinese company in 2017, they must be using a different sensor system.  The first Kinect for Windows device as of 2015  Kinect for Windows v2 was also discontinued in 2015  Kinect for Xbox One was discontinued in October 2017 Microsoft reportedly confirmed that it has ended production for the Kinect’s USB adapter 1/2018 There are used Kinect units sold at thrift stores, I have seen batches of unknown condition units sold in bulk on Ebay (often the rotater gears are stripped) .  Leaving one to obtain the software.  Kinect is back but different The new Kinect targeting Internet of Things applications is the Azure Kinect using updat...

Real Time Video Processing FPGA board

Image
DOCS PYNQ is an open-source project from Xilinx ®   that makes it easy to design embedded systems with Xilinx Zynq ®  Systems on Chips (SoCs).  Using the Python language and libraries, designers can exploit the benefits of programmable logic and microprocessors in Zynq to build more capable and exciting embedded systems.  The System on Chip has both an ARM processor based Linux system and a Xilinx Zynq FPGA. PYNQ uses CPython which is written in C, and integrates thousands of C libraries and can be extended with optimized code written in C. The PYNQ-Z1 board is an academic version of a Xilinx evaluation board, it is almost identical to the standard product, but the board is colored pink, and had been offered with academic discounts.  It has HDMI input and output, the available libraries include OpenCV, but there are only 2 functions that have been translated to the FPGA, that run at full frame rate.  One could tackle creating an overlay to ...

New High Performance (Risc-V) Processors Available to DIYers

Image
  A new processor that has been in the works for a few years is the open source Risk-V architecture, developed at Berkeley.   Enthusiastic participation by the industry has fostered the development of various chip designs and open source software, Chinese companies in particular love the idea of not having to pay a license fee on each chip they manufacture, the one big loser is the maker of Arm processors, used in almost every cell phone and 32 bit hobby computer.  Western Digital, a big Arm user has been a strong sponsor of Risk-V and has committed to making a Billion Risk-V processors.  I don't have much info on the technical advantage of this design, they do have a small instruction set, which simplifies optimizing hardware, this new entry should cause an overall drop in processor prices. A simplified evolutionary timeline: 3/8/14 EE Times Article By Berkeley Designers, RISC-V: An Open Standard for SoCs - The case for an open ISA Founded in 2015, the RI...

Compact Arduino Speech Production, ( Speak n Spell )

Image
  A clever codes reverse engineered the speech compression technology Texas Instruments devised in the late 70s, and created Arduino code called Talkie to play word sets lifted from old product ROMs, there are 6 different libraries included in the examples totaling about 1000 words with redundancies, there's also experimental compression of a 2 minute Susan Vega song that occupies 24k of space.. The file links below are all Arduino IDE format. The original code version works properly with the timers in the Arduino Mega 168 or 328 running at 16 Mhz. This version is tweaked to work on an 8Mhsz processor . Here's a version translated for ATmega32U4 processor . This version was translated by the creator of the Teensy computer boards . One collector of vintage voice synthesis files created a $95 music synthesizer plugin product, chipspeech based on these robotic voices. You can listen to tracks from an album incorporating these sounds . Speak n Spell uses a rat...