4th Week: CST 334

 

Topic: Memory Virtualization

Hello everyone,

    In this week’s material we covered the idea of memory virtualization in detail like the ideas of free space management, paging, translations, page tables, swapping mechanism and swapping polices. One important thing I learned was understanding how memory is allocated dynamically using APIs for example malloc. This allows programs to allocate memory from the heap. The goals of memory allocators are to balance accuracy, speed, and efficiency while managing memory effectively.

    But dynamic memory allocation causes an issue known as external fragmentation. This happens when free memory is divided into small, noncontiguous blocks that are too small to use effectively. A solution to the problem of external fragmentation is paging, which eliminates the issue by dividing virtual memory into fixed-sized units called pages. These pages are mapped to physical memory, which is divided into page frames, and a page table is used to translate virtual addresses to physical addresses.

    Paging gives advantages such as flexibility because it eliminates the requirement for address spaces to be contiguous and avoids external fragmentation. Also, paging simplifies free space management by allocating memory in fixed size chunks, pages, and maintaining a free list. However, one drawback of this technique is that it may cause internal fragmentation where pages are only partially filled leading to unused memory within each page.

    We also looked into making the page table translations faster, using Translation Lookaside Buffer (TLB), a cache that stores virtual address (VA) to physical address (PA) mappings. When translating a VA to a PA, the Memory Management Unit (MMU) first checks the TLB. A TLB hit is a hit if is found in the cache, the MMU retrieves the PA directly from the TLB, saving the overhead of accessing the page table in main memory. But on a TLB miss, the MMU must do a page table lookup and if the page is not found in memory, then it is considered a page fault. A page fault requires the operating system to load the page into main memory from secondary storage which is expensive. This process involves page swapping from disk. When a page fault occurs, the operating system sets a trap, causing the CPU to switch to kernel mode to handle the disk fetching process. A lot of information!

Comments

Post a Comment

Popular posts from this blog

CST438 Software Engineering: Week 8 (Final week)

 One of the most important things I learned is that software engineering is about much more than just writing code. While coding is a big part, there's a strong emphasis on documentation, understanding client needs, and getting continuous feedback. It’s about designing systems so others can easily improve them and applying principles like "low coupling and high cohesion" to keep things modular and maintainable.      Another key lesson was understanding the entire development process from end to end. This was the first time I worked with both the front end and back end together, fully grasping the HTTP protocol (including how methods like POST work), how JSON is used, and how the front end communicates with the back end. Seeing how everything fits from database interactions to deploying the app online—gave me a fascinating glimpse into the full workflow and a lot of practical experience.      Another important lesson was learning how to work as part of ...
Read more

CST438 Software Engineering: Week 1

What did you expect a course in Software Engineering would cover? I expected it to cover how to build applications from the ground up and the design behind application development. I thought we would look at other things we’ve been learning in this program, like database management, front end development, and back end work, and then put it all together into one complete project. After completing week 1, how has your opinion changed? Well, I never really thought much about keeping an application’s design maintainable for future improvements or changes. There is a lot more that goes into software engineering than just simple programming. It is  about planning ahead, documenting your work, checking for errors, and testing properly to avoid issues later. I’ve also realized software engineering looks at the big picture it is not just about coding and design but about how software impacts people and how teams work together to build it. Developers need to stay motivated, think critical...
Read more

CST370 Week 5

Hello everyone,  This week we studied binary‑tree traversals preorder, inorder, and postorder but I kept mixing them up. I finally got it when I attended TA office hours, and it clicked. Walking through Quick Sort step by step helped me see how the Decrease and Conquer approach works by breaking problems into smaller chunks. Then we looked at Kahn’s algorithm, and watching it remove nodes with no incoming edges reminded me of picking a pivot in Quick Sort. Seeing these patterns repeat helped me realize that many algorithms work by simplifying things piece by piece. 
Read more