FIT1047 Introduction to computer systems

计算机组成和体系结构【4】MARIE机器 计算机组成和体系结构【4】MARIE机器MARIE = the Machine Architecture that is really Intuitive and Easy一、 寄存器1. 累加寄存器 Accumulator ,AC2. 存储器地址 寄存器 Memory Address Register,MAR3. 存储器数据缓存 寄存器 Memory Buffer Register,MBR4. 程序计数器 Program Counter,PC5. 指令 寄存器Instruction Register,IR 阅读详情


FACULTY OF INFORMATION TECHNOLOGY
FIT1047 Introduction to computer systems, networks and security – S2 2024
Assignment 2 – Processes and MARIE Programming
      Purpose
   Your task
Part 1a: Submit your reflections. See details below.
Part 1b: Write a short report describing the processes running on your computer. Part 1c: Disassemble and add comments to a MARIE program.
Part 1d: Write a MARIE program that can display bitmap numbers.
Part 2: Interview for Part 1d.
        Value
    Word Limit
    Due Date
        Assessment Criteria
Part 1a is assessed based on relevance of the submission to the unit.
Part 1b is assessed based on correctness and completeness of the descriptions. Part 1c is assessed based on correctness of the code and the labels/comments. Part 1d is assessed based on correctness of the code, as well as the documentation/comments.
Part 2 is assessed based on the understanding of the code you have written. See instructions for details.
    Processes and programs are what make computers do what we want them to do. In the first part of this assignment, students will investigate the processes running on their computers. The second part is about programming in MARIE assembly language. This will allow students to demonstrate their comprehension of the fundamental way a processor works.
The assignment relates to Unit Learning Outcomes 2, 3 and 4.
  25% of your total marks for the unit.
The assignment is marked out of 60 marks.
See individual instructions.
Part 1a-1d: 11:55 pm Friday 30 August 2024 Part 2: Interview during Week 8
   Submission
   • This is an individual assignment (group work is not permitted).
• Turnitin and MOSS will be used for similarity checking of all submissions.
Ignore the Turnitin warning or error messages for zip files or other non-
document files.
• DRAFT submission is not assessed. DRAFT upload confirmation email from Turnitin is not a submission. You must press the submit button to accept the terms and conditions in Moodle.
• In this assessment, you must not use generative artificial intelligence (AI) to generate any materials or content in relation to the assessment task.
• You will need to explain and extend your code in an interview. (Part 2)
 
INSTRUCTIONS
This assignment has five parts. Make sure you read the instructions carefully.
For Part 1a, collect your reflections from each week’s Ed Lesson and create a single PDF document. You can simply copy/paste your reflection, but please add headings for each week. Submit the PDF through the Moodle Assignment activity.
Part 1b-1d, you need to submit files through the Moodle Assignment activity.
Files that are required via Moodle Assignment Submission:
• Part 1a: one pdf file (containing reflection from Week 4, 5 and 6) • Part 1b: one pdf file (containing answer)
• Part 1c: one .mas file for Part 1c
• Part 1d: one .mas file for Part 1d
You may see the following error message from Turnitin. You can ignore this error message!
FACULTY OF INFORMATION TECHNOLOGY
     Late Penalties
• 5% deduction per calendar day or part thereof for up to one week
• Submissions more than 7 calendar days after the due date will receive a mark
of zero (0) and no assessment feedback will be provided.
       Support Resources
     Feedback
Feedback will be provided on student work via:
• general cohort performance
• specific student feedback ten working days post submission
    See Moodle Assessment page
 Part 2 is an interview in Week 8. Instructions will be available in Moodle and communicated via an announcement post.

FACULTY OF INFORMATION TECHNOLOGY
Part 1a: Reflections (Hurdle – you MUST submit it in order to pass this assignment!)
Complete your reflection activities in Week 4 to Week 6 Ed Lesson and copy/paste them into a pdf file. Each week the reflection must have at least 100 words (relevant and meaningful to the specific week).
Failure to submit all relevant week’s reflections (missing all submissions or incomplete submissions) will result in your assignment 1 having a maximum mark of 49% only. For example, if the overall combined mark is 61/100, it will be scaled to 49/100. If the overall combined mark is 44/100 then it will remain as 44/100.
You may use the template provided in Moodle to write down your reflection.
Part 1b: Processes (10 marks total)
For this task, write a brief report about processes that you observe running on your computer. You can use one of the following tools (depending on your operating system):
• On Windows, use the Task Manager
• On macOS, use the Activity Monitor
• On Linux, use a command line tool like htop, top, or the ps command
Answer the following questions:
1. Briefly describe the columns displayed by the tool you use that relate to a) memory usage and b) CPU usage of a process. What can you say about the overall memory usage of all processes, compared to the RAM installed in your computer? Include graphs or charts for the comparison. (5 marks)
2. Pick a process you perhaps don’t know much about, or which you did not expect to find running on your computer. Try to find out and describe briefly what it does. (5 marks)
Include a screenshot of your processes in the report along with CPU/memory usage graphs and/or charts. The screenshot should show between 5 and 10 processes.
The word limit for this part (both questions together) is 500 words (about 1 page, not including images and tables). Use APA 7th referencing style in the report.
Submit your report for this part (Part 1b) as a PDF file (independent of the other parts) in Moodle.
 
Part 1c: MARIE Disassembly (20 marks total)
Follow the link on Moodle to access your personalised MARIE memory screenshot for this task.
Important: Your memory screenshot is different from the one other students are working on. Only download the file while you are correctly logged into Moodle with your own student account.
Task 1: Disassemble the memory (10 marks)
Based on the memory contents, recreate the MARIE program that corresponds to your personalised memory screenshot. This is called “disassembling” the machine code, since it is the opposite operation of “assembling” the MARIE code into the binary memory contents.
For each memory cell, decode the instruction and (if applicable) the address that the memory cell is encoding. You can make the following assumptions:
- There is exactly one Halt instruction in the code
- Every memory location after the Halt instruction contains data
- Any memory location that contains the value 0 is data (even before the Halt instruction)
Here is an example of a memory screenshot and the corresponding decoded MARIE program: Disassembled program:
Input
Add 005
Output
Jump 000
Halt
DEC 10
Note: You need to decode the actual instructions. E.g. for the first memory location, HEX 5000 would not be a valid answer. The contents of all memory that follows the Halt instruction is considered to be data. Therefore, DEC 10 is the correct decoding of location 5 (instead of JnS 00A), and HEX 00A would also be correct. You don’t need to list all the locations containing zeros starting from address 006 (these will be filled with zeros by the assembler anyway).
Tip: You can verify that your disassembled code is correct by entering it into the MARIE simulator, assembling it and comparing the memory contents to the screenshot you started from.
FACULTY OF INFORMATION TECHNOLOGY
  
Task 2: Add labels (5 marks)
Now update the program you decoded in Task 2.1. Removing all hard-coded memory addresses by adding labels to replace all memory locations that are used as addresses in the program instructions. Labels should have meaningful names in the context of what the program does (i.e., not just A, B, C). For the example above, this could result in the following program:
MainLoop, Input
Add Ten
Output
Jump MainLoop
Halt
Ten, DEC 10
Task 3: Add comments (5 marks)
Comment the code based on your understanding of what it does. Comments should describe the function of the different parts. E.g., if you identify a subroutine in the code, add a comment at the start of the subroutine that describes what it does, and whether it takes any arguments.
For this part (Part 1c), you need to submit one .mas file containing your final code. Do not submit one .mas file per each subtask!
FACULTY OF INFORMATION TECHNOLOGY
 
Part 1d: MARIE Programming (22 marks)
In this task you will develop a MARIE application that draws numbers on the screen. We will break it down into steps for you.
Each task requires you to write code and documentation. On Moodle, you will find a template for the code. Your submission must be based on this template, i.e., you must add implementations of your own subroutines into the template. The template already contains the main program that calls the subroutines.
Your code must contain readable comments and meaningful labels for your tutor / marker to understand the logic flow of your program (e.g. the purpose of a subroutine, jump / skipcond statement etc.).
Interview (Part 2): You will be required to join an interview to demonstrate your code to your tutor in week 8 (after the submission deadline). Failure to demonstrate will lead to zero marks being awarded for the entire Part 1d, regardless of your submission in Moodle. In addition, during the interview (Part 2), you will also need to answer further questions about your submitted code (see below for details).
Code similarity: We use tools such as MOSS and Turnitin to check for collaboration and copying between students. If you copy parts of your code from other students, or you let them copy parts of your code, this will result in a report to the Academic Integrity team. As a result, you may receive a penalty such as 0 marks for the entire assignment, 0 marks for the whole unit, or in severe cases (such as contract cheating), suspension or expulsion from Monash University.
Rubric: The marking rubric on Moodle provides details for the marking. A correctly working MARIE program that covers all tasks and is well documented will receive full marks. Missing/incomplete documentation will result in a loss of up to 1⁄4 of the task’s marks.
Introduction: Bit-mapped displays
So far, the only output capability we have seen in the MARIE system is using the Output instruction, which will print a single 16-bit value. Many computers of course are capable of displaying arbitrary graphics, often in high resolution and great colour depth.
In the lectures on input/output systems, we have seen that one way to implement this is to map a certain location of the memory to an output device. I.e., writing to that memory location (e.g. using a Store instruction) causes the output to happen.
In the simplest form of graphics hardware, we can dedicate part of the RAM to be graphics memory. Each memory cell corresponds to a pixel on screen, and the value in the memory cell encodes the colour of the pixel. That way, we can create arbitrary graphics by simply writing values into the memory.
FACULTY OF INFORMATION TECHNOLOGY
 
FACULTY OF INFORMATION TECHNOLOGY
The MARIE simulator has a feature called Display, which you access from the list of tabs that also shows the output log, RTL log etc:
  The display shows the memory from address F00 to address FFF as a 16x16 pixel screen. The value in the memory locations represents the colour of the pixels. We will only use the colours black, represented as 0, and white, represented as FFFF. When you start the MARIE simulator and assemble your code, the memory starting from location F00 is (usually) filled with zeroes, which means that the display is black. Let’s now change the contents of the memory using some Store instructions:
Load White
Store 0F80
Store 0F81
Store 0F82
Store 0F83
Halt
White, HEX FFFF
After running this program, the display will look like this:
 You can see that the first four pixels in the 9th row have now turned white.

Task 1: Clearing the display (4 marks)
Write a subroutine SubClearDisplay that turns all pixels in the graphics memory white. Remember that the graphics memory ranges from address 0F00 to address 0FFF, and that white pixels are represented by the value FFFF. Document your subroutine with comments.
Task 2: Painting a number (10 marks)
The template for this task contains data for bitmaps of the digits 0-9, stored at the label Font. Each digit consists of 3x5 pixels of data. The first 3 words are the first row of pixels, the next 3 words are the second row, and so on. For example, the digit 2 is represented as
0 0 FFFF FFFF FFFF 0 FFFF 0 FFFF 0 FFFF FFFF 000
You can see the pattern here, the zeros “paint” the shape of the character 2 in black, with the background in white (FFFF).
Your task is to write a subroutine called SubPaintDigit that paints a digit into the graphics memory. The start of the subroutine needs to look like this:
PaintDigitCharacter, HEX 0
PaintDigitDisplay, HEX 0
SubPaintDigit, HEX 0
In the PaintDigitCharacter argument, we pass the address of the first pixel data in the font for the digit we want to paint. In the PaintDigitDisplay argument, we pass the address of the top- left corner where we want to start painting in the graphics memory. For example, to paint the digit 0, starting from the second pixel in the second row, we could use the following code:
    Load FontAddr
    Store PaintDigitCharacter
    Load Display22
    Store PaintDigitDisplay
    JnS SubPaintDigit
    Halt
    Display22, HEX 0F11
Note that the address 0F11 (label Display22) lies exactly 17 words after the start of the graphics memory. This means we’re skipping the first row (16 words) and the first pixel in the second row (1 word).
Here we simply use FontAddr to refer to the first character (for the digit 0). For the other characters, we would have to add a corresponding offset into the font memory.
FACULTY OF INFORMATION TECHNOLOGY
 
FACULTY OF INFORMATION TECHNOLOGY
In order to paint a digit in your subroutine, you can follow this “recipe”:
- Your subroutine should contain two nested loops.
- Each digit contains 15 pixels, so you need to loop through those 15 pixels, load each one from
the font definition and store it into the graphics memory. This is the outer loop of your
subroutine.
- After each set of 3 pixels, you need to start in the next row of the graphics display. This means
that if you were currently writing into graphics memory at address X, you now need to continue writing at address X plus the width of the display minus the width of a character. This is the inner loop of your subroutine.
- Once you have “copied” all 15 pixels from the font definition into the graphics memory, you can exit the subroutine.
Your subroutine needs to contain sufficient comments to enable someone else (like the person marking your assignment) to understand the purpose of each line of your code.
Task 3: Counting down (8 marks)
Your final task is to implement a subroutine SubCountDown that clears the screen and then counts down from 9 to 0, drawing those digits on the bit-mapped display using the subroutines developed in the previous tasks.
In order to get full marks, your code needs to use a loop that decrements a counter and calls SubPaintDigit based on the value of the counter, rather than a sequence of instructions that calls SubPaintDigit with each digit’s address. Use additional subroutines to structure your code nicely.
You will notice that it would be nice for the countdown to wait for a fraction of a second between digits. Think of a way you can achieve this, so that the countdown takes (more or less) exactly 10 seconds on your computer to execute. Document how you achieved this in the code comments.
A sample output video for this part (part 1d) is provided for your reference:
https://drive.google.com/file/d/1NQPMNCuLoVxf7dD8omFqacS9SRCcefQC/view?usp=drive_link
For this part (Part 1d), you need to submit one .mas file, based on the template, containing the code for all subroutines. Do not submit one .mas file per each subtask!
  
Part 2: Interview (8 marks)
You need to demonstrate the code you submitted for Part 1d Task 1–3 to your tutor in an interview (to be conducted in Week 8) after the submission deadline. Failure to explain how your code works will result in 0 marks for the individual tasks that you cannot demonstrate.
In addition, you will be asked to modify the code you submitted in certain ways and explain how the MARIE concepts work that you were required to use for the individual tasks. These additional questions add up to 8 marks for this task (Task 2).
Failure to attend the interview will result in 0 marks for the entire Part 1d and 2, regardless of your submission in Moodle.
Submission Checklist: Please check before you submit: Altogether, you should submit the following files:
• Part 1a:
o 1 pdf file for Week 4 to Week 6 Reflection
Failure to submit this file will result in getting only a maximum
49% marks in this assignment (combined part 1 and part 2)!
• Part 1b:
o 1 pdf file – your answers to Part 1b questions
• Part 1b:
o 1 .mas file – only one .mas file containing your final code. Do NOT
submit one .mas file per each subtask! You MUST NOT submit any
pdf file in this part!
• Part 1d:
o 1 .mas file – only one .mas file containing your final code based on the template, containing the code for all subroutines. Do NOT submit one .mas file per each subtask! You MUST NOT submit any pdf file in this part!
Failure to submit the correct file type (e.g. submit a word file instead of pdf file in Part 1b, or a pdf file in Part 1c or Part 1d), or missing any file in your submission (e.g. your weekly reflection) will result in significant marks deduction.
Please check before you submit! You CANNOT revert any submission!!!
(Note: Draft submission is not assessed!)
FACULTY OF INFORMATION TECHNOLOGY
 

《BiLSTM with Multi-Polarity Orthogonal Attention for Implicit Sentiment Analysis》论文阅读笔记 文章名:《BiLSTM with Multi-Polarity Orthogonal Attention for Implicit Sentiment Analysis》 作者Jiyao Wei a,1, Jian Liaoa,1, Zhenfei Yanga, Suge Wanga,b,∗, Qiang Zhaoa 年份 2019 1 Introduction 目前文本情感分析多分析带... 阅读详情

相关推荐

别再为淘宝滑块发愁了!一个Python脚本自动获取x5sec cookie的保姆级教程

本文提供了一套完整的Python自动化解决方案,帮助开发者轻松应对淘宝滑块验证码问题。通过模拟浏览器行为、精准识别滑块缺口位置和生成拟人化移动轨迹,脚本能够自动获取关键的x5sec cookie,显著提升爬虫效率。教程涵盖环境配置、验证机制分析、代码实现及反检测策略,是解决淘宝滑块验证的实用指南。

weixin_30300225的博客 577

FIT1047 java /Python

PurposeYour taskValueWord LimitDue DateSubmissionFeedbackInputAdd 005OutputJump 000HaltDEC 10Add TenOutputHalt。

wolfminy的博客 817

从按键开机到I2C隔离:手把手拆解一个智能硬件项目里的MOS管实战配置

本文详细解析了智能硬件项目中MOS管的实战配置,从按键开机电路到I2C总线隔离,全面覆盖电源管理和信号隔离的关键技术。通过PMOS+NMOS组合方案实现可靠的开机功能,并采用双NMOS方案优化I2C通信隔离,提升系统稳定性和信号质量。

weixin_28691741的博客 326

四川大学网安计算机组成和体系结构期末整理

Chapter 1(教材第一章):Introduction 1.3 The Common Measures § Kilo (千)- (K) = 1 thousand = 10^3 and 2^10 § Mega (兆)- (M) = 1 million = 10^6 and 2^20 § Giga (鲒)- (G) = 1 billion = 10^9 and 2^30 § Tera (汰)- (T) = 1 trillion = 10^12 and 2^40 § Peta (帕)- § = 1

POEMlion的博客 3013

实现自动打包py及替换pyinstaller --add-data参数的方法

实现自动打包py及替换pyinstaller --add-data参数的方法

isS_MK_GmbH的博客 3468

opencv基础57-模板匹配cv2.matchTemplate()->(目标检测、图像识别、特征提取)

OpenCV 提供了模板匹配(Template Matching)的功能,它允许你在图像中寻找特定模板(小图像)在目标图像中的匹配位置。模板匹配在计算机视觉中用于目标检测、图像识别、特征提取等领域。: 首先,加载目标图像和要匹配的模板图像。: 选择适当的匹配方法,例如 cv2.TM_CCOEFF、cv2.TM_CCOEFF_NORMED、cv2.TM_CCORR、cv2.TM_CCORR_NORMED、cv2.TM_SQDIFF 或 cv2.TM_SQDIFF_NORMED。每种方法对应不同的匹配计算方式。

小海聊智造 1万+

FIT1047 Introduction to computer systems, networks and security

InputAdd TenOutputHaltmapgraphicsmemorypixelcolourDisplayLoad WhiteStore 0F80Store 0F81Store 0F82Store 0F83Halt0 0 0HaltXX。

vx_codinghelp的博客 1188

AI大模型实战——模型测评:如何评估大模型的表现

大模型评测体系解析:从数据集到基准测试 本文系统梳理了大模型评测的关键要素。在数据集方面,重点介绍了C-Eval等中文评估套件,强调数据质量需通过人工验证保证。评测维度涵盖自然语言理解、知识推理、代码生成等核心能力。基准测试作为标准化评估方法,Glue、SuperClue等平台提供了公平比较的依据。文章指出厂商自评榜单存在水分问题,建议关注第三方评测。同时解析了少样本和零样本两种prompt模式的应用场景差异。最后解释了"SOTA"这一术语在描述模型最优性能时的应用价值。整体呈现了大模型

小志的博客 898

FIT1047 Introduction to computer systems networks and security - S2 2024 2R

Purposedowhatwewantthemtodo.processesisaboutinMARIEtheirrelates toUnitYour taskPartbelow.Part1b:reprunningon yourcomputer.Part1c:MARIEprogram.Part1d:Write aMARIEnumbers.Part 2:

wx_codinghelp的博客 1286

【C++】小猫钓鱼纸牌游戏--C++实现

小猫钓鱼问题: 描述:初始,两个人各有n张牌(牌面0-10),顺序固定,只能按顺序出牌。然后,两个人轮流在桌面上以接龙的方式出牌,当某个人放置的牌在桌面上已经出现时,他收起相同牌面及其中间的牌作为自己的筹码,同时需要再放一张牌到桌面。某人手上的牌出完后,另一个人继续出,直到两人手上都没牌为止,输出结束时两个人手上的筹码数。 #include<iostream> using namespace std; struct queue { int head; //队列首 int tail; //

qq_41344430的博客 3202

杰理蓝牙AC696X开发实战:自定义蓝牙名称与提示音配置指南

本文详细介绍了杰理AC696X蓝牙芯片开发中自定义蓝牙名称与提示音配置的完整实战指南。文章从SDK工具定位入手,逐步讲解如何通过配置工具修改蓝牙名称,并重点阐述了替换提示音的正确步骤与音频文件(如WAV格式、16kHz采样率)的详细参数要求,最后指导如何生成并验证最终固件,帮助开发者快速实现个性化设备定制。

weixin_29161719的博客 562

FIT1047 Introduction to computer systems networks and security – S2 2024 3 – NetworksWeb

Purposeethatit,identifyitspotentialissues. Sneedto analyseandprotocolsused.isrelated toUnitYour taskPartPart1b: Yourepotasks. TheinPart 2:beprovidedlater inMoodle.)Value30Part1;20%

wx___codinghelp的博客 874

RTX6000 PRO BLACKWELL 最优 CUDA 版本选择指南(适配 vLLM/SGLang 大模型推理)

本文针对NVIDIA Blackwell架构RTX6000PRO显卡部署vLLM/SGLang大模型推理引擎时的CUDA版本选择问题,提出专业建议。分析表明,CUDA12.9(需搭配552.06+驱动)是最优选择,因其完全支持BW架构特性,且与vLLM/SGLang最新版本深度适配。文章提供详细的Linux环境部署教程,包括驱动升级、CUDA12.9安装及vLLM/SGLang配置步骤,并指出CUDA13.0目前适配不完善,不推荐新手使用。最后给出常见问题解决方案,强调版本匹配对推理性能和稳定性的重要性。

2501_93651177的博客 682

FIT1047 Introduction to computer systems networks and security – S2 2024 4 – CybersecurityPython

Part 2a:Part 2b:one itemrecorda videooneof the3news●lessthanthanleastonlycontainoniswecannot verifyifyourisdonebyanotherperson).Ifyouhavetechnicalonturningonthe MonasheSolution。

hometime2的博客 1531

温度PID控制PLC程序

温度PID稳定控制程序,适用于欧姆龙PLC,对其他PLC也具有很大参考价值

Statistics and Machine Learning Toolbox

Statistics and Machine Learning Toolbox(带书签),MATLAB统计学习工具

FIT1047 Introduction to computer systems networks and security – S2 2024 1Python

Purposecoreof allcomputerIn thisknowledgeofandsimplifyrelates toUnit1 and 2.Your taskPartbelow.Partin thebelow.Yous wellasLogisimfilescircuits.In-persSeeMoodledetails.Valuee belowforhow。

miraitowaty的博客 1043

假人有限元模型

假人模型,车辆碰撞假人有限元模型,适合做乘员安全的研究者利用

学习C#开源报表组件Seal Report(45:报表服务安全配置)

学习Seal Server Manager软件内与报表服务安全配置相关的各类属性的用途

gc_2299的博客 161

霍尼韦尔SoftMaster使用说明

霍尼韦尔SoftMaster使用说明

Advance Steel工具栏菜单的创建

创建jigdesignpalettedata.xml文件并放置于...\INT\Steel\Support\Toolbars\目录下,随后修改C:\ProgramData\Autodesk\Advance Steel 2025\INT\ASSettings_User.xml,添加对应键值对以加载该配置。示例中设置面板位置、大小及可见性,通过ParentId指定父窗口,确保自定义工具栏正确显示。

Advance Steel二次开发 40

C# 系列教程_12 字符串 string 保姆级系统教程(面向新手,大量示例 + 原理讲解)

C#中string与String为同一类型,均为不可变引用类型,任何“修改”操作均返回新字符串。常用创建方式包括字面量、逐字字符串@""、插值$""及new String()。字符串本质是字符数组,支持索引访问但不可修改。所有方法如IndexOf、Substring、Replace等均不改变原字符串,返回新实例。

zhanghaha1314的博客 306

Swift 字符详解:从 Character 到字符串遍历

Swift中Character代表扩展字形簇,符合人类对“一个字符”的直观理解。它与String为集合关系,可通过for-in遍历或数组拼接。Character基于Unicode,组合字符(如带重音字母、表情符号)视为整体,count统计字形簇数而非码点或字节。访问需用String.Index,不可直接整数下标。字符支持比较、大小写转换及ASCII互转,但必须用双引号声明,且与String类型区分。理解其设计可避免常见错误,提升字符串处理可靠性。

2402_85743155的博客 307

C#转Python第5.3篇:类型提示与静态分析:Python 的“伪静态类型“

Python的类型提示是“伪静态类型”——注解由IDE和工具(如mypy)使用,解释器运行时忽略。与C#强制编译时类型检查不同,Python类型注解仅作开发辅助,提升代码可读性、智能提示与静态分析能力。通过List[int]、Optional[str]、Protocol等特性,实现类型安全与泛型支持。Python 3.14起引入延迟求值与annotationlib,优化性能并支持前向引用。核心理念:类型注解是“给开发者看的文档”,而非运行时保障。

hanzhixintianxia的博客 267

第6章第4节:自定义分词器

本节详解Elasticsearch自定义分词器,解析analyzer三段式结构(char_filter→tokenizer→filter),掌握常用组件如ik_max_word、synonym_graph、ngram等。通过_analyze API调试分词效果,实现“索引时扩展同义词、搜索时不带”的生产级方案,提升中文搜索与智能提示体验。

啊渊的专栏 798

C# 数据结构和特殊类型

.NET Core(现称.NET)提供丰富数据结构,优化性能与开发效率。日常开发首选List<T>和Dictionary<TKey,TValue>;高性能场景推荐Span<T>与Memory<T>减少内存分配;并发环境使用ConcurrentDictionary等线程安全集合;避免非泛型集合如ArrayList。

cccloveforever的博客 765

SQL 基础查询与模糊查询全解 + C# WinForm 数据库连接实战:ADO.NET、参数化查询与 where 1=1 动态条件

上篇把表建好、关系连好。,再。从select *到,从到,从拼字符串到参数化,最后讲清楚where 1=1到底解决什么问题——顺便把课上那道选择题的答案讲透。

2501_94699871的博客 172

C# NModbus4 实战:ModbusTCP 读取以太网温湿度传感器

本文基于C#与NModbus4库,实战演示如何通过ModbusTCP协议读取以太网温湿度传感器(如H-THRJ45)数据。涵盖设备配置、寄存器地址解析、缩放换算及常见调试坑点,提供可运行的控制台Demo,适用于机房动环、GSP库房等物联网场景。代码支持异步扩展、数据持久化与告警逻辑,助力快速搭建自研采集系统。

weixin_41865733的博客 284

C# 系列教程_10 方法(Methods)保姆级精讲

本文系统讲解C#方法的定义与使用,涵盖语法结构、参数传递、返回值、作用域及重载机制。通过无参无返回值、有参、有返回值等示例,阐明方法如何减少冗余、提升代码可读性与维护性。重点强调方法重载(同名异参)、返回值类型不可单独决定重载、局部变量作用域限制等核心概念,并提醒新手常见误区,如缺少return、误判重载条件等。最后介绍可选参数与传值机制,帮助初学者掌握方法调用的规范与技巧。

zhanghaha1314的博客 340

BPE分词器的完整实现:字节对合并与词表构建

本文深入解析了大语言模型分词器的核心机制,以字节级BPE为例,揭示其如何将文本转化为token。通过字节编码实现无外生词(OOV)问题,但带来跨语言价差与计数故障等挑战。文章完整实现BPE训练与编码逻辑,强调预切分正则与语料配比的关键作用,并指出分词器不仅是技术组件,更是决定模型感知粒度与成本结构的底层决策——它塑造了模型对文字的理解方式,也影响着每一次推理的效率与费用。

2601_96733043的博客 293
上一篇: C167 microcontroller
下一篇: FINC5090 - Finance in the Global Economy
vx_codinghelp
博客等级 码龄2年 543粉丝 68原创
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
 条评论被折叠 查看
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付:10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值