Assignment 5:  Dynamic Storage Allocator

If you want to use slip days you have to tell us before the deadline (by emailing Sudhanshu Gupta and/or Rongcui Dong). We won't by default assume slip days for late submissions, and won't retroactively apply slip days.

No email submissions will be accepted.

If you code does not compile on the CSUG machine, you get a 0.

Introduction

In this lab you will be writing a dynamic storage allocator for C programs, i.e., your own version of the malloc, free, and realloc routines.  You are encouraged to explore the design space creatively and implement an allocator that is correct, efficient and fast. 

DUE DATES:

11:59pm, Saturday, May 8. 

Hand Out Instructions

Start by copying /u/cs252/labs21/malloclab/malloclab-handout.tar to a protected directory in which you plan to do your work.  Then give the command: 

    tar xvf malloclab-handout.tar
This will cause a bunch of files to be unpacked into the directory.  The only file you will be modifying and handing in is mm.c

The mdriver.c program is a driver program that allows you to evaluate the performance of your solution.  Use the command make to generate the driver code and run it with the command ./mdriver -V.  (The -V flag displays helpful summary information.)

Looking at the file mm.c you'll notice a C structure team into which you should insert the requested identifying information about the one or two individuals comprising your programming team.  Do this right away so you don't forget. 

How to Work on the Lab

Your dynamic storage allocator will consist of the following four functions, which are declared in mm.h and defined in mm.c

  • int mm_init(void);
  • void *mm_malloc(size_t size);
  • void mm_free(void *ptr);
  • void *mm_realloc(void *ptr, size_t size);

The mm.c file we have given you implements the simplest but still functionally correct malloc package that we could think of.  Using this as a starting place, modify these functions (and possibly define other private static functions), so that they obey the following semantics: 

  • mm_init:  Before calling mm_malloc, mm_realloc, or mm_free, the application program (i.e., the trace-driven driver program that you will use to evaluate your implementation) calls mm_init to perform any necessary initializations, such as allocating the initial heap area.  The return value should be −1 if there was a problem in performing the initialization, 0 otherwise. 

  • mm_malloc:  The mm_malloc routine returns a pointer to an allocated block payload of at least size bytes.  The entire allocated block should lie within the heap region and should not overlap with any other allocated chunk. 

    We will compare your implementation to the version of malloc supplied in the standard C library (libc).  Since the libc malloc always returns payload pointers that are aligned to an 8 byte boundary, your malloc implementation should do likewise. 

  • mm_free:  The mm_free routine frees the block pointed to by ptr.  It returns nothing.  This routine is only guaranteed to work when the passed pointer (ptr) was returned by an earlier call to mm_malloc or mm_realloc and has not yet been freed. 

  • mm_realloc:  The mm_realloc routine returns a pointer to an allocated region of at least size bytes with the following constraints: 

    • if ptr is NULL, the call is equivalent to mm_malloc(size);

    • if size is equal to zero, the call is equivalent to mm_free(ptr);

    • if ptr is not NULL, it must have been returned by an earlier call to mm_malloc or mm_realloc.  The call to mm_realloc changes the size of the memory block pointed to by ptr (the old block) to size bytes and returns the address of the new block.  Notice that the address of the new block might be the same as the old block, or it might be different, depending on your implementation, the amount of internal fragmentation in the old block, and the size of the realloc request. 

      The contents of the new block are the same as those of the old block, up to the minimum of the old and new sizes.  Everything else is uninitialized.  For example, if the old block is 8 bytes and the new block is 12 bytes, then the first 8 bytes of the new block are identical to the first 8 bytes of the old block and the last 4 bytes are uninitialized (garbage).  Similarly, if the old block is 8 bytes and the new block is 4 bytes, then the contents of the new block are identical to the first 4 bytes of the old block. 

These semantics match the the semantics of the corresponding libc malloc, realloc, and free routines.  Type man malloc to the shell for complete documentation. 

Heap Consistency Checker

Dynamic memory allocators are notoriously tricky beasts to program correctly and efficiently.  They are difficult to program correctly because they involve a lot of untyped pointer manipulation.  You will find it very helpful to write a heap checker that scans the heap and checks it for consistency. 

Some examples of what a heap checker might check are: 

  • Is every block in the free list marked as free? 
  • Are there any contiguous free blocks that somehow escaped coalescing? 
  • Is every free block actually in the free list? 
  • Do the pointers in the free list point to valid free blocks? 
  • Do any allocated blocks overlap? 
  • Do the pointers in a heap block point to valid heap addresses? 

Your heap checker will consist of the function int mm_check(void) in mm.c.  It will check any invariants or consistency conditions you consider prudent.  It returns a nonzero value if and only if your heap is consistent.  You are not limited to the listed suggestions nor are you required to check all of them.  You are encouraged to print out error messages when mm_check fails. 

This consistency checker is for your own debugging during development.  When you submit mm.c, make sure to remove any calls to mm_check as they will slow down your throughput.  Style points will be given for your mm_check function.  Make sure to put in comments and document what you are checking. 

Support Routines

The memlib.c package simulates the OS portion of the memory system for your dynamic memory allocator.  You can invoke the following functions in memlib.c

  • void *mem_sbrk(int incr):  Expands the heap by incr bytes, where incr is a positive non-zero integer and returns a generic pointer to the first byte of the newly allocated heap area.  The semantics are based on the Unix sbrk function, with two notable exceptions: 
    • the built-in sbrk returns -1 on an allocation error, mem_sbrk returns NULL
    • mem_sbrk accepts only a positive non-zero integer argument
  • void *mem_heap_lo(void):  Returns a generic pointer to the first byte in the heap. 
  • void *mem_heap_hi(void):  Returns a generic pointer to the last byte in the heap. 
  • size_t mem_heapsize(void):  Returns the current size of the heap in bytes. 
  • size_t mem_pagesize(void):  Returns the system's page size in bytes (4K on Linux systems). 

The Trace-driven Driver Program

The driver program mdriver.c in the malloclab-handout.tar distribution tests your mm.c package for correctness, space utilization, and throughput.  The driver program is controlled by a set of trace files.  Some small traces are included in the malloclab-handout.tar distribution.  The larger traces that we will test your file with are located at /u/cs252/labs21/malloclab/malloctraces/.  These larger traces are automatically run if you execute mdriver without a -f argument.  Each trace file contains a sequence of allocate, reallocate, and free directions that instruct the driver to call your mm_malloc, mm_realloc, and mm_free routines in some sequence. 

The driver mdriver.c accepts the following command line arguments: 

  • -t tracedir:  Look for the default trace files in directory tracedir instead of the default directory defined in config.h
  • -f tracefile:  Use one particular tracefile for testing instead of the default set of tracefiles. 
  • -h:  Print a summary of the command line arguments. 
  • -l:  Run and measure libc malloc in addition to the student's malloc package. 
  • -v:  Verbose output.  Print a performance breakdown for each tracefile in a compact table. 
  • -V:  More verbose output.  Prints additional diagnostic information as each trace file is processed.  Useful during debugging for determining which trace file is causing your malloc package to fail. 

Programming Rules

  • You should not change any of the interfaces in mm.c
  • You should not invoke any memory-management related library calls or system calls.  This excludes the use of malloc, calloc, free, realloc, sbrk, brk or any variants of these calls in your code. 
  • You are not allowed to define any global or static compound data structures such as arrays, structs, trees, or lists in your mm.c program.  However, you are allowed to declare global scalar variables such as integers, floats, and pointers in mm.c
  • For consistency with the libc malloc package, which returns blocks aligned on 8-byte boundaries, your allocator must always return pointers that are aligned to 8-byte boundaries.  The driver will enforce this requirement for you. 

Evaluation

You will receive zero points if you break any of the rules or your code is buggy and crashes the driver.  Otherwise, your grade will be calculated as follows: 

  • Correctness (20 points).  You will receive full points if your solution passes the correctness tests performed by the driver program.  You will receive partial credit for each correct trace. 

  • Performance (35 points).  Two performance metrics will be used to evaluate your solution: 

    • Space utilization:  The peak ratio between the aggregate amount of memory used by the driver (i.e., allocated via mm_malloc or mm_realloc but not yet freed via mm_free) and the size of the heap used by your allocator.  The optimal (unachievable) ratio would be 1.  You should find good policies to minimize fragmentation in order to make this ratio as close as possible to the optimal. 
    • Throughput:  The average number of operations completed per second. 

    The driver program summarizes the performance of your allocator by computing a performance index, P, which is a weighted sum of the space utilization and throughput

    P = wU + (1−w) min (1, T / Tlibc)

    where U is your space utilization, T is your throughput, and Tlibc is the estimated throughput of libc malloc on your system on the default traces.  (The value for Tlibc is a constant in the driver (4000 Kops/s) that the TA established when he configured the release).  The performance index favors space utilization over throughput, with a default of w = 0.6

    Observing that both memory and CPU cycles are expensive system resources, we adopt this formula to encourage balanced optimization of both memory utilization and throughput.  Ideally, the performance index will reach P = w + (1−w) = 1 or 100%.  Since each metric will contribute at most w and 1−w to the performance index, respectively, you should not go to extremes to optimize either the memory utilization or the throughput only.  To receive a good score, you must achieve a balance between utilization and throughput.  You will receive full performance score if your performance index P is above 0.95.

  • Style (10 points). 

    • Your code should be decomposed into functions and use as few global variables as possible. 
    • Your code should begin with a header comment that describes the structure of your free and allocated blocks, the organization of the free list, and how your allocator manipulates the free list. 
    • Each subroutine should have a header comment that describes what it does and how it does it. 
    • Your heap consistency checker mm_check should be thorough and well documented. 

    You will be awarded 5 points for a good heap consistency checker and 5 points for good program structure and comments. 

Hints

  • Use the mdriver -f option.  During initial development, using tiny trace files will simplify debugging and testing.  We have included two such trace files (short{1,2}-bal.rep) that you can use for initial debugging.  Other trace files are located in /u/cs252/labs21/malloclab/malloctraces/. 

  • Use the mdriver -v and -V options.  The -v option will give you a detailed summary for each trace file.  The -V will also indicate when each trace file is read, which will help you isolate errors. 

  • Compile with gcc -g and use a debugger.  A debugger will help you isolate and identify out of bounds memory references. 

  • Understand every line of the malloc implementation in the textbook.  The textbook has a detailed example of a simple allocator based on an implicit free list.  Use this is a point of departure.  Don't start working on your allocator until you understand everything about the simple implicit list allocator. 

  • Encapsulate your pointer arithmetic in C preprocessor macros or gcc in-line functions.  Pointer arithmetic in memory managers is confusing and error-prone because of all the casting that is necessary.  You can reduce the complexity significantly by writing macros for your pointer operations.  See the text for examples. 

  • Do your implementation in stages.  The first 9 traces contain requests to malloc and free.  The last 2 traces contain requests for realloc, malloc, and free.  We recommend that you start by getting your malloc and free routines working correctly and efficiently on the first 9 traces.  Only then should you turn your attention to the realloc implementation.  For starters, build realloc on top of your existing malloc and free implementations.  But to get really good performance, you will need to build a stand-alone realloc

  • Use a profiler.  You may find the gprof tool helpful for optimizing performance. 

  • Start early!  It is possible to write an efficient malloc package with a few pages of code.  However, we can guarantee that it will be some of the most difficult and sophisticated code you have written so far in your career.  So start early, and good luck! 

Turn In Instructions

The main assignment will be submitted using the script /u/cs252/bin/TURN_IN. 

    /u/cs252/bin/TURN_IN .

Before running the TURN_IN script, be sure that you have

  • included your full name and email address in the comment at the top of mm.c
  • removed any extraneous print statements
  • included any appropriate commentary on your code in a separate README file or as C comments in mm.c