University of California at Berkeley
College of Engineering
Department of Electrical Engineering and Computer Science

EECS 61C, Spring 2008

Project 4

Due Thursday, May 1, 2008 @ 11:59:59pm

TA in charge: Keaton Mowery (kmowery@berkeley.edu)


Notes/Updates

Last Update: 2008-05-01 18:30


Contents


Brief Introduction

This project will give you intimate knowledge of cache logic through implementation of, you guessed it, cache logic. We will be providing you a MIPS simulator called TIPS (Thousands of Instructions Per Second), that will be able to run MIPS instructions. However, the cache logic is broken (read: conveniently non-existant)!

Your job is to implement the cache logic behind the cache so that TIPS can make use of the many benefits cacheing entails. You may choose to complete this project either by yourself or with one other partner.

If you choose to do this project individually, your slip days will work as normal. However, if you choose to do this project with another person, the group's total slip days will be the average of the two team members' days, rounded down. For example, if both partners have 2 slip days, then the team will have 2 slip days. If one partner has 1 slip day and the other has 3, the team will have 2 slip days.

An oracle will be provided soon so you can compare your solution with ours.


Setup on Inst Machines


GUI Walkthrough

The GUI has been designed to be straight-forward. There are four main components to the GUI interface: register display, execution log, cache display, and control panel.

 

A description of each of the GUI widgets are described as follows:

  1. Register display -- detailed view of the current state of the registers
  2. Execution log -- log of actions by TIPS. Messages can be displayed in this box using the append_log() function.
  3. Cache display -- current snapshot of the state of the cache. The meaning of the column headings on each unit are:
    • Blk - block number
    • V - valid bit
    • D - dirty bit
    • LRU - LRU data
    • Tag - Tag for the block
    • Numbers (00, 04, etc.) - offset in the cache block data
  4. Config Cache -- configure the cache parameters
  5. Load Program -- loads a dump file for execution
  6. Step -- execute one instruction
  7. Run -- automate execution
  8. CPU -- reset the PC and reinitialize registers
  9. Cache -- flush the cache
  10. Output -- clear the execution log
  11. Quit -- exit TIPS

There is also a text-based version of the GUI for those who prefer it. You can access it with the following call:

  $ ./tips -nogui

Type help at the TIPS prompt to get a list of commands usable in this mode.


Your Task

To complete this project, you must complete the accessMemory() function in cachelogic.c. This function will handle accessing actual memory, using the accessDRAM() function. Thus, the behavior of accessMemory() function will be a cache that will call the accessDRAM() function as needed.

To ensure a variety of caches can be simulated, you will be required to dynamically support 5 properties:

More information about the variables you will be working with and the functions at your disposal can be ascertained by looking over tips.h.

You should keep the following things in mind when formulating the code:


Getting Started

Look over tips.h and cachelogic.c. tips.h gives you an overview of how the cache simulator is put together. A section of that file has been marked as important, so read it to get an idea of what functions and variables you will be using. cachelogic.c contains a slightly more detailed explanation of what you will be writing in the accessMemory() function.

The cache data structure is divided into three levels:

There are two methods to access the LRU information of a block. The first method is via lru.value, a field that will hold LRU information in integer format. The other method is via lru.data, a field that will hold LRU information represented in another format (its type is void*, which means it can be a pointer to anything). You are free to use either method to represent the LRU information, so long as the LRU behaves in a deterministic fashion (i.e. no two valid blocks will ever be candidates for replacement at the same time).


Organization of the memory functions in TIPS

In a nutshell, the accessMemory() function is a function that acts as a communication layer between the CPU and DRAM. The code within the accessMemory() function manipulates the cache data structure defined in tips.h.

All your code MUST be contained in cachelogic.c - specifically in accessMemory() and possibly the LRU functions (depending on how you implement the LRU algorithm). You may add helper functions as long as you do not modify any code outside of cachelogic.c. Do not change any file other than cachelogic.c. Do not change the prototypes of existing functions. Just to be clear, modify only the body of the given functions, adding helper functions if needed.


Creating Dump Files for Testing

TIPS contains a subset of the MIPS R2000 processor instruction set. As a result, it accepts MIPS code assembled into binary. The dump file can be loaded into the program by clicking on the "Load Program" button.

Dump files can be created in several ways. If you are working on the inst machines, you can use the script included in the project files:

	./assemble_mips input.s output.dump

Working from home is a bit trickier, but still perfectly doable. First, download Mars 3.4.1 from here. Once you have the jar file, you have two options. If you prefer to use the GUI, go ahead and start Mars and open your assembly file. Ensure that the "Delayed Branching" setting is enabled in the Settings menu. Now, assemble the source file. Then, select the "Dump Memory" option from the File menu. Ensure that the ".text" memory segment is selected and that the dump format is Binary, and you should be good to go. If you prefer to use the command line, you can do something like this:

	java -jar Mars-3.4.jar a db dump .text Binary output.dump input.s


Questions to Answer

Answer the following questions in the README file. You can safely assume that all associative caches in these questions use an LRU replacement policy.

Question 1:

For this question, assume 1KiB of addressable memory, 64B of L1 cache with 8B blocks. You will probably find it useful to draw the cache as you work through this problem.

The code in question:

	#define ARRAY_SIZE 64
int total = 0;
char * array = malloc(ARRAY_SIZE * sizeof(char)); /* Line 1 */ // Note: chars are 1-byte wide
for(int x = 0; x < ARRAY_SIZE; x++) array[x] = x; /* Line 2 */
for(int x = 0; x < ARRAY_SIZE; x++) total += array[x]; /* Line 3 */
for(int x = 0; x < ARRAY_SIZE; x++) total += array[x]; /* Line 4 */

a. Assume the cache is direct-mapped and array is a pointer to the memory at offset 0b1001111000, what is the ratio of cache hits to cache misses while Line 4 is being executed?

b. Assume the cache is fully associative and array is a pointer to the memory at offset 0b1001111000, what is the ratio of cache hits to cache misses while Line 4 is being executed?

c. Assume the cache is direct-mapped and array is a pointer to the memory at offset 0b1001111101, what is the ratio of cache hits to cache misses while Line 4 is being executed?

d. Assume the cache is fully associative and array is a pointer to the memory at offset 0b1001111101, what is the ratio of cache hits to cache misses while Line 4 is being executed?

e. Do these results surprise you? Why or why not? Explain what's going on. What would happen if the cache were 2-way set associative? 4-way?

Question 2:

Describe a situation in which a 2-way set associative cache would out-perform a fully associative cache. Vice-versa? You should describe these situations in relation to the attributes of the cache. (i.e. accesses are made to fill every block of the cache, hitting memory block-size apart each time...) Otherwise, the examples can be as contrived as you wish.

Question 3:

Explain the differences between write-through and write-back caches, and when/why one might be preferred over the other.

Question 4:

Suppose for a moment that virtual memory was included in this project. How would you expect the input to accessMemory() to change? (in other words, what type of information should the cache now receive in order to do its thing?) YOU ARE NOT EXPECTED TO IMPLEMENT THIS.

Question 5:

Briefly describe the changes you would have to make to your implementations in this project to implement an L2 cache. Assume you're given a second cache struct. How would your L1 cache implementation change? How, if at all, would your L2 cache implementation need to be different from your original cache implementation? YOU ARE NOT EXPECTED TO IMPLEMENT THIS.


Submitting Your Solution

Repeated for emphasis: All your code MUST be contained in cachelogic.c (this is the only code you submit)

In testing your project, we will use all of our own files (including the Makefile) EXCEPT for your submitted cachelogic.c.

A README file should be submitted with your answers to the questions above and anything the grader should know reader about your project.

Only one person per group will need to submit; the submission process will ask for your partner's login. If both members of the group submits, the last submission will be the one we grade.


Bugs and Suggestions

If you think you have found a bug in the TIPS GUI/oracle/code, please let the staff (i.e. the TA in charge) know via email or the newsgroup. Likewise, feel free to let us know of any suggestions you have. Bugs and suggestions will be addressed in updated files for this project if necessary.


Frequently Asked Questions

Q: What are the definitions of associativity and set?
A: Associativity is the number of cache blocks that a given memory block can be placed using the memory block's index.

Set is the collection of blocks associated with the SAME index. The number of unique indexes is equal to the number of sets you have.

Q: Is an oracle available for my home computer?
A: Oracles will be forthcoming for Cygwin/Windows, OS X, and Linux in the next few days.

Q: I am updating the cache in my function, but the GUI isn't doing anything. What is going on?
A: The GUI needs to be informed when changes are made to the cache. The GUI functions at your disposal are mentioned in tips.h.

Q: How much of the GUI behavior do I have to copy?
A: You will only be graded on your implementation of the cache logic in cachelogic.c. You do not have to include the GUI update function calls if you don't want to, but they will help you visually understand how caching works.

Q: Should I follow the LRU algorithm seen in COD?
A: No, you should not follow the LRU algorithm in COD. The LRU algorithm in COD is an approximation LRU algorithm, which produces ties between valid blocks that are candidates for replacement. As a result, it is suggested you should use a true LRU algorithm, where there is a unique ranking between valid blocks. In other words, there should never be any randomness in your LRU replacement policy, nor should it always default to kicking just one block out in the event of a "tie" (two LRU values being the same for two valid blocks -- blocks with valid bit set to 1).

Q: Do I have to use numbers for my LRU algorithm?
A: No, you do not have to use numbers for your LRU algorithm. If you want to use something else (such as time structs), use lru.data as the pointer to this data. Make sure you update the init_lru() function and the lru_to_string() function.

Q: Are the addresses in this project byte addressed or word addressed?
A: They are byte addressed.

Q: What do you mean by unit of transfer between cache and physical memory is in blocks?
A: This means if you have blocks in your cache with a size of 4 words, you must always move data to and from physical memory in units of 4 words. There should never be a time in which the amount of data you move to physical memory from cache be less or greater than the block size of the cache.

Q: Can I add more files to the project?
A: You can, but we will not be able to test your solution in the autograder. We will ignore all files you submit except for cachelogic.c (and the README, of course). Do NOT put any code needed by your cachelogic.c (including header files and such) in any external file. If that's not clear, edit only cachelogic.c please -- your grade will appreciate not being a 0.

Q: Can I work from home without the GUI?
A: Yes, there is a "No GUI" mode for TIPS that can be activated by adding the -nogui flag on the command line when starting TIPS. For instructions, type "help" on the prompt after starting up TIPS.

Q: Can I work from home without using SSH?
A: Yes, the Makefile in the proj4 folder will allow compilation on Cygwin/Windows, OS X, and Linux automatically.

Q: Can I start TIPS with a dump file already loaded?
A: Yes, TIPS will interpret the last argument on the command line as a name of a file. If you want to start up TIPS with a dump file named test23.dump, you can type the following:

  % ./tips test23.dump

Q: I typed in make and it says there is an error in the Makefile. What's wrong?
A: Use gmake

Q: I ran gmake and I can't compile this project on Cygwin/Windows because of *some reason*. What do I have to do to get it to compile on Cygwin/Windows?
A: This project's GUI interface is built on
GTK+. Compiling this project on Cygwin/Windows requires you at least install the packages mentioned in "Setup on Windows Machines (Cygwin)." Information about specifics of the general GTK+ installation can be found here.

Q: Can I use other compilers on Windows?
A: All the project files are configured for use on Sparc/Solaris or Cygwin/Windows. If you want to use other compilers (i.e. Microsoft Visual Studio), you will have to find the libraries usable with those compilers and make the appropriate modifications. Remember, you must make sure your program compiles on the lab computers in 271 Soda.


Setup on Windows Machines (Cygwin)

The Cygwin Oracle (version 2 - outdated!) can be found here. It contains a bug pertaining to block replacement... It should be very obvious if you run into it. It will only run from within Cygwin, so make sure that you put it in your Cygwin home directory. You may have to make it executable before it will run:

$ chmod +x tips_oracle_cygwin
$ ./tips_oracle_cygwin


Setup on OS X

Leopard users: Go download XCode 3.0 from Apple's site. You'll have to register as a developer, but accounts are free. After downloading the 1.1GB XCode dmg, open it and install XCode. Then, continue with installing MacPorts below.

Non-Leopard users (10.4 and before): Go download XCode 2.5 from Apple's site. Again, you may have to register as a developer. Then, download the X11 package and install it, then continue with installing MacPorts below.

Next, install MacPorts (instructions are on their site), and enter these commands in Terminal:

$ sudo port install gmake
$ sudo port install gtk2

You should be good to go! Please let me know if you have any issues; I may have forgotten a step or two.

The Oracle (version 3) for OS X can be found here. You may have to make it executable before it will run:

$ chmod +x tips_oracle_osx
$ ./tips_oracle_osx


Setup on Linux

If you've made it this far in the semester with Linux, you should have the necessary compilers and libraries needed to compile the project. However, to ensure that you have the correct GUI libraries, you might need to install a few new packages.

If you are using a Debian based distribution (like Ubuntu):

$ sudo aptitude update && sudo aptitude install build-essential libgtk2.0-dev

If you're still getting errors, you may need to install another package. Look at the output of "uname -a" and look for a string like "Linux 2.6.15-28-686". Take note of the bolded numbers (your arch) and do:

$ sudo aptitude install linux-headers-arch

If you're on a different distribution, look in your package management system for the appropriate packages.. Email me if you have issues.

The Oracle (version 3) for Linux can be found here. You may have to make it executable before it will run:

$ chmod +x tips_oracle_linux
$ ./tips_oracle_linux

Setup on Other Machines

If you're using something other than Windows, OS X or Linux, you're on your own. Let us know if there's a problem with which we can help.