OSU eCampus CS271 Computer Architecture and Assembly Language review & recap

This post is part of an ongoing series recapping my experience in Oregon State University’s eCampus (online) post-baccalaureate Computer Science degree program. You can learn more about the program here.

Six-word summary: Where my 4.0 GPA went to die

CS271 is a pedantic deep-dive into your computer’s inner workings and an opportunity to enjoy coding in a highly restrictive language. There’s special emphasis on going step-by-step through the stack frame and understanding what’s in every register and even some much-appreciated training on “what order to do things in” when building a program.

CS271 Review

CS271 often felt like it was made of two distinct halves that ran in parallel: the first half is made up of the cute little MASM programs you write for the homework that’s due every other week. The other half is the mountains of computer architecture trivia and methods for calculating various things you learn for the bi-weekly exams, such as binary/hex conversions, Hamming codes, memory usage, optimization, recursive output, stack frame changes, the step-by-step changes in registers as a program runs or steps through a loop, and so on.

The class is very well organized. They tell you precisely what to do for the homework and what the output should look like, line by line. Unlike many classes in OSU’s CS program thus far, you won’t need to do a ton of Googling to learn the material – the lectures and the book reading are actually adequate preparation for most of the class’s homework and quiz questions. The workload is consistent. Each week follows the same pattern, and the homework’s complexity ramps up slowly.

However, the homework (in which you write some Assembly language code) is rather light and may give you a false sense of security come quiz and exam time. The true heavy hitters in this class are the self-checks you do after the lectures, the weekly exercises, and the bi-weekly quizzes. Mastering the material found in all of these is essential for surviving the midterm and the final.

Whether 271’s material is interesting is going to depend on your own likes and dislikes. I thought most of it was squarely in the “hey, that’s neat” category, but I thought it focused too much over low-level details that (probably) won’t directly enhance my career in software. (And that’s just like, my opinion, man – I’m sure some people who take 271 find their calling and never resurface from the depths of CPU architecture, binary math, hex addresses, and bit ordering.)

Class structure

  • 10 weeks
  • Multiple briskly-paced weekly lectures
  • An ungraded/untimed “self-check” exercise worksheet for every single lecture – you should actually do these, the quizzes and exams pull from them
  • A weekly exercise quiz with a 6 hour time limit – you get two attempts and they keep the best score, but each attempt has a different batch of questions so it is advised to take it twice to see both question banks
  • 4 bi-weekly tests, 1-hour time limit, each worth 2.5% of your final grade
  • 6 Assembly-language coding projects (homework assignments) total – each one gives you two full weeks to work on it, with the grading rubric released at the start of the 2nd week
  • Proctored midterm, proctored final
  • Steady, even workload
  • Less self-teaching than some OSU classes (still quite a bit of rote memorization and practice, though)
  • NO GROUP PROJECTS!

Tips for CS271

This was my first OSU CS class where I couldn’t quite earn a solid A (I got an A-). It certainly wasn’t for lack of effort or time put in – I think I cleared 20 hours of practice most weeks and I worked on this class’s stuff daily.

There’s a ton of material and the tests are very picky – they’re not multiple choice where half the answers are obviously silly, they’re fill-in-the-blank or multiple choice where all the options seem like they could be right. (The tests are also almost half your final grade. Ouch.)

When I look back on what I could’ve done better and what I would advise future students, these are the things that come to mind.

CS271 tip #1: Get the book, it doesn’t suck!

This is one of the books that’s worth buying. It’s approachable with many  examples to follow and, surprisingly, many quiz questions are lifted line-for-line from the book.

See Assembly Language for x86 Processors on Amazon.com

I liked having a physical book because it freed up screen space during quizzes and while working on homework. I found it helpful to read on paper, practice in a notebook, and really work problems over and over until I had the process down cold. (That’s just my personal preference; I seem to retain more if I read and work on paper.)

CS271 tip #2: Make yourself some worksheets and grade yourself

Speaking of rote memorization and brute-force practice, I highly recommend making yourself some worksheets out of the weekly self-checks and bi-weekly quizzes. Print them out and work through them (over and over) until you can do all of them quickly and perfectly.

This technique was time-consuming but it went a long way to cementing the steps to solving all the various types of CS271 problems. As I mentioned earlier, the homework is borderline useless for the exams. The exams draw their questions almost entirely from the weekly self-checks and bi-weekly quizzes.

The exams tend to layer multiple concepts together into one question and they’re autograded, not graded by a potentially generous TA, so your goal for CS271’s tests should be to become very fast at solving every kind of problem you encounter in the practice materials.

CS271 tip #3: Get a binary/hex calculator

The Casio fx-115ES can do math in binary and hex: addition, subtraction, and conversions between binary/hex and back again.

I had no idea this kind of calculator even existed and the course materials make no mention of it. Frustratingly, I only found out about it through the class’s Slack chat. I was doing conversions by hand like a total chump, which would have cost me dearly on the midterm either in terms of time spent or stupid mistakes or both. The hex and binary questions on the exam were made trivial by having this calculator. I don’t know why the syllabus doesn’t open with “GET THIS CALCULATOR”.

As an alternative to the physical calculator, the course lets you use your operating system’s built-in calculator. (I did this course’s work on my PC but I take tests via ProctorU on Macbook since that’s where my webcam is, so I preferred the physical calculator over having to get familiar with multiple OS calculators.)

CS271 tip #4: Pack that cheat sheet to the density of a neutron star

I made mine in Google Docs, which will let you pick a font size as small as 6 (you have to type it in, it’s not in the dropdown).

One side of my CS271 cheat sheet – the back side was all code examples.

I don’t care for exam cheat sheets. I think allowing cheat sheets drives students to put their efforts into guessing what details should be on the sheet, which sets the student up for an inevitable wave of frustration and disappointment when they realize 2 seconds into the exam that they guessed wrong. Letting students bring a cheat sheet seems to inspire instructors feel free to ask all kinds of minutia-focused questions instead of testing for a more general understanding of the concepts.

But… nobody asked my opinion, and I doubt the cheat sheets are going away any time soon, so my advice here is to pack your sheet as full as you can get it. I think a good cheat sheet for CS271’s midterm and final would be at least 50% code examples, which is what the back side of my sheet was full of.

CS271 tip #5: The extra credit on the homework was generally a waste of time for me

I tried to do the extra credit (of which there are 1-3 points up for grabs on every bi-weekly assignment), but for the most part I found the extra credit opportunities frustrating and almost as time-consuming as the rest of the assignment itself. Your mileage may vary, but I should’ve spent the time on worksheet drills instead.

CS271 tip #6: If I could do one thing differently…

I’d spend more time studying all the nuances of what happens in the system stack while a program runs a loop. Here’s me practicing on a whiteboard, but despite all this preparation I still think this is where I lost points on the exam. I was probably off by 4 on some memory address or miscounted loops somewhere.

And that’s it for CS271! I’m glad I’m done with it; I’ve heard it’s one of the tougher courses and for me, that was definitely true. I worked very hard in this class and I advise all future students to do the same. The tests are no joke and your ability to track and manage tiny details will be tested like never before.

Next up: CS290 – Web Development. See you all next quarter!

How to convert a 4-byte hexadecimal sequence in a little-endian architecture into decimal

Here is my technique for solving the problem of converting a 4-byte hexadecimal sequence in a little-endian architecture into decimal.

This may seem rather niche but it was a surprisingly large part of Week 5 in my CS271 class. The class’s materials and extra help I found around the web seemed to go off on tangents that were interesting but unrelated to solving this kind of problem quickly, and all I really wanted was a simple step-by-step guide I could use on exams.

Therefore, I am sharing my simple 5-step technique for converting hex to decimal in a little-endian architecture here! Hope someone else finds it helpful.

Example problem:

The four-byte sequence 0x86 0x65 0x53 0x82 stored in consecutive memory cells in a little-endian architecture represents ___________ (decimal) when interpreted as a 32-bit signed integer.

From reading this, we know: 

  • it’s little-endian, so we are going to reverse the order of the bits
  • our result will be signed

Step 1: Reverse the bytes

Take the bits in blocks of two and work right to left.

0x86 0x65 0x53 0x82 becomes 0x82536586

Step 2: Look at the most significant bit and determine if there will be a negative result

0x82536586   <-- that's this dude in red here

The most significant bit here contains an 8.

We know that in hex a most-significant bit of 7 or more means we are looking at a negative number. (If this were a positive number, ie: the most significant bit is between 0 and 6, then skip ahead to Step 4.)

Step 3: Since we are working with a negative number, flip the bits (subtract our hex sequence from FFFFFFFF) and add 1

  FFFFFFFF
- 82536586
  7DAC9A79

Add one to the result:

7DAC9A79
      +1
7DAC9A7A

The result is the hex sequence we will use for the next step.

7DAC9A7A

Step 4: Multiply each term by 16 raised to a power

To convert a hex value into a decimal value, we multiply each “position” in the hex sequence by 16 raised to a power. Working from right to left, we know that furthest-right position is 16^0 (so, just a 1). The second-from-right position is 16^1 (so, just a 16). The third-from-right position is 16^2, and so on.

Recall that A = 10, B = 11, C = 12, D = 13, E = 14, and F = 15 when working in hex.

(7*16^7)+(D*16^6)+(A*16^5)+(C*16^4)+(9*16^3)+(A*16^2)+(7*16^1)+A

The result of this long sequence is:

2108463738

Remember that negative sign? Now’s a good time to stick it on.

-2108463738

And there you have it: the result is -2108463738.

Some final notes

  • Be sure to observe whether the problem expects a signed decimal result or an unsigned decimal result. If the problem is asking for unsigned, you can skip the FFFFFFFF subtraction step entirely, even if the most significant bit is 7 or higher.
  • Remember that when working with a signed hexadecimal number, you look at the most significant bit to determine if it’s negative or positive.
    • 0-7 = positive
    • 8-F = negative
  • If you had to do the flipping step, don’t forget to put that negative sign onto your final answer!