Machine problem 6: Malloc Solution

   Objectives

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



   Implementation Details

The only file you will be modifying in your repository 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.)

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 comparing 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 8 bytes, your
    malloc implementation should do likewise and always return 8-byte
    aligned pointers.

-   `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
        `ptr` 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. 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. Refer to the
[malloc manpage](http://linux.die.net/man/3/malloc) 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.

   Support Routines

Functions in "memlib.c" simulate 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
    identical to the Unix `sbrk` function, except that `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 in "mdriver.c" tests your "mm.c" package for correctness, space
utilization, and throughput. The driver program is controlled by a set of *trace
files*, each of which 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 and the trace files are the same
ones we will use when we grade your submitted "mm.c" file.

`mdriver` 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 allocate any global or `static` *compound data
    structures* such as arrays, structs, trees, or lists in your code. However,
    you are allowed to declare global *scalar variables* such as integers, floats,
    and pointers.

-   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.

-   While you may certainly refer to the book's implicit list based implementation
    for inspiration or help, you **may not plagiarize it**!

   Grading

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 (35 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 (45 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 ratio equals to 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 (*U*) and throughput (*T*):
    
    *P* = *w* &times; *U* + (1 - *w*) &times; min(1, *T*/*T_libc*)
    
    where *T_libc* is the estimated average throughput of `libc`
    malloc on your system on the default traces. The value for
    *T_libc* is a constant in the driver (10000 Kops/s), based on a
    conservative estimate of `libc` malloc throughput on the server. 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. 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.
    
    While the ideal value for your performance index is 1 (100%), a score of
    0.9 (90%) or higher will net you the full 45 points for performance. A
    lower index will be used to compute your score according to the formula
    min(45, (*P* + (1-0.9)) &times; 45).



   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.

-   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.

-   Use `gdb`! 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.

-   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`.

-   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 right away, and good luck!