Embedded System & FPGA

NEORV32

https://github.com/stnolting/neorv32

Probabilistic Models of Cognition

https://probmods.org/

A Browsable Petascale Reconstruction of the Human Cortex

ProbNum Python Lib

最近接觸到這個新的概念 叫做Probabilistic Numeric

感覺十分有趣

我的直覺是跟大腦的運作似乎有些關聯

因為大腦是透過建構模型而不是紀錄來與世界互動

而PN好像也是這樣在做數學上的運算

從積分 線性代數到解微分方程式

都可以帶入PN的概念

產生的計算結果 也比一般傳統的數值分析

多了機率與統計的分析

BMM Summer Course 2021

Cutting Edge: The next wave: Probabilistic programming with Stuart Russell

Probabilistic Machine Learning by Philipp Hennig at the University of Tübingen

https://www.youtube.com/playlist?list=PL05umP7R6ij1tHaOFY96m5uX3J21a6yNd

DDS with Poisson Distribution Spike is Fun

Udemy: Docker and Kubernetes

懶的看書 買了Udemy 的課程 

Micro Service 背後的確很多 Moving parts

EKS -> Kubernetes -> Docker (Docker Hub) -> Your Code

(EKS) Provide Nodes and Volume ->

(Kubernetes) Management Pods, Load Balance and Volume ->

(Docker) Generate Images and Host Images ->

Your Code 

....

Why Micro Service in the 1st places.

Poetry another dependency management for Python

Poetry is a tool for dependency management and packaging in Python.

It allows you to declare the libraries your project depends on and it will manage (install/update) them for you.

https://python-poetry.org/docs/

OpenFPGA

OpenFPGA aims to be an open-source framework that enables rapid prototyping of customizable FPGA architectures

What Is Causal Inference?

QBayLogic (Clash Maintainer)

https://qbaylogic.com/blogs/all/

RIOT powers the Internet of Things like Linux powers the Internet.

RIOT powers the Internet of Things like Linux powers the Internet.

重構的思考

搭公車時 忽然對之前要重構的問題 有了一個想法

問題是這樣

幾個IO HAL 與 中介的資料結構 就說是Queue 吧

IO HAL 很直接的卡在資料結構程式中

因此在使用上很不方便

修改的方式已經都知道了

但是問題是在我想重構的過程

我一直在想如何一次把整個IO HAL 與資料結構 拆解掉

因為牽扯很多其他的部分

因此一直遲遲不敢動

今天在公車上想到方法了

其實很簡單

就是

一個IO 一個IO 搬就好了

花了好幾個星期才有這個領悟

實在是.....

這就是設計與修改與重構的不同

設計雖然是一步一步來

但是因為空白依舊很多 因此修正時很大膽

重構則因為系統互相牽制以及一般都缺乏測試

因此想要大膽的修正是完全錯誤的行動思考

只有一步一步來才是正確的

另外

800行的三層If then else

該用甚麼方式面對呢

 

Retrocomputing with Clash

Clash這個Haskell DSL 語言引導我 認識了

Haskell

Functional Programming

RISC-V

Chisel

Scala

Rust

哈哈一段4年有趣的旅程

Clean Code - Uncle Bob (6)

Clean Code - Uncle Bob (5)

Clean Code - Uncle Bob (4)

Clean Code - Uncle Bob (3)

1. We don't ship shit
2. We will always be Ready
3. Stable Productivity
Same design speed as start
4. Inexpensive adaptability
5. Continuously Improvement
6. Fearless competences (with Test)
7. We are not dump on qa
Qa donot find anything
8. Test Automatio
9. We cover for each other , Team work
10. Introduction to "Honest Estimates"

Clean Code - Uncle Bob (3)

Clean Code - Uncle Bob (2)

Clean Code - Uncle Bob

Every programmer should watch.

視角問題

之前寫程式 遇到一個問題

就是

收到資料了 該如何通知 需要的人

簡單的方法就是直接把那個人帶在身邊 笨

或是弄一個即時作業系統用 通知 煩

想想想

後來

幹麻通知人家 他不會來問喔

你寫的程式 對接手的人 是資產還負債 這才是問題 ＠＠

MICROPY_HW_CLK_USE_HSI = (1) --> Pay attention to following setting

/**
* @brief System Clock Configuration
*
* The system Clock is configured for F4/F7 as follows:
* (HSx should be read as HSE or HSI depending on the value of MICROPY_HW_CLK_USE_HSI)
* System Clock source = PLL (HSx)
* SYSCLK(Hz) = 168000000
* HCLK(Hz) = 168000000
* AHB Prescaler = 1
* APB1 Prescaler = 4
* APB2 Prescaler = 2
* HSx Frequency(Hz) = HSx_VALUE
* PLL_M = HSx_VALUE/1000000
* PLL_N = 336
* PLL_P = 4
* PLL_Q = 7
* VDD(V) = 3.3
* Main regulator output voltage = Scale1 mode
* Flash Latency(WS) = 5
*
* The system Clock is configured for L4 as follows:
* System Clock source = PLL (MSI)
* SYSCLK(Hz) = 80000000
* HCLK(Hz) = 80000000
* AHB Prescaler = 1
* APB1 Prescaler = 1
* APB2 Prescaler = 1
* MSI Frequency(Hz) = MSI_VALUE (4000000)
* LSE Frequency(Hz) = 32768
* PLL_M = 1
* PLL_N = 40
* PLL_P = 7
* PLL_Q = 2
* PLL_R = 2 <= This is the source for SysClk, not as on F4/7 PLL_P
* Flash Latency(WS) = 4

Asynchronous programming in MicroPython :)

https://github.com/peterhinch/micropython-async

All you need to learn about LoRa and LoRaWan

AWS Service 2021

想用 AWS IOT Core Lorawan 很久沒上 AWS 

發現 服務又變多了

Protothread

http://dunkels.com/adam/pt/

Main features:

• Very small RAM overhead - only two bytes per protothread and no extra stacks
• Highly portable - the protothreads library is 100% pure C and no architecture specific assembly code
• Can be used with or without an OS
• Provides blocking wait without full multi-threading or stack-switching
• Freely available under a BSD-like open source license

Example applications:

• Memory constrained systems
• Event-driven protocol stacks
• Small embedded systems
• Sensor network nodes
• Portable C applications

Simple is Beauty