- Review linked_list.h and implement those functions in linked_list.cc
- Ensure that your code compiles with no warnings nor errors.
- Pass functional tests by running the function test suite. You can do this by invoking 'make run_functional_tests'.
- Pass valgrind checks to ensure your code does not leak memory. You can do this by invoking 'make run_valgrind_tests'
The C++ class linked_list is declared for you in linked_list.h. It consists of an inner class node, and the linked_list class is composed of node objects.
In terms of memory allocation and deallocation, your library depends on user-specified malloc() and free() functions, registered via the linked_list::register_malloc() and linked_list::register_free(). functions. See the malloc_fptr and free_fptr function pointer variables in linked_list.cc, which you can use in your implementation of your linked list. One of the reasons for this is code like instrumented_malloc() in linked_list_test_program.c. Another reason is to allow you to specify to me whether you'd like to use a memory allocation library other than libc's implementation of malloc()/free().
We don't ask you to implement an iterator in the C++ version, although you're welcome to do it and if you're a paid participant we'll gladly review it for you.
Our philosophy is to try to make this look as "C like" as possible to keep the two versions of the challenge equivalent, but using excellent C++ style coding conventions, because that's what a professional C++ repository demands.
It's a lot of work for little benefit, and in our view is beyond the scope of what we're trying to do with Pointer Wars. Remember this is a high quality C++ coding challenge, which means we expect that iterators inherit from appropriate classes such that any C++ idiom that would use the iterator "just works". That's a high bar.
You'll also note in particular the lack of copy constructor, move constructor, and assignment operator overloads. Intentional. Our code never does those operations, so we're not going to ask you to implement them. That's just paying a "C++" tax over the C participants, but we will say that if this was a professional code repository, it is extremely unlikely that a linked list without that functionality would fail code review.
Send us email (pointerwars2025@gmail.com) with a link to your GitHub page, or attach files to the email. If your code passes functional tests and valgrind tests, we'll send you code feedback and performance numbers. Level 1 performance numbers are microbenchmarks that measure down to the nanonsecond how fast your linked list operations are.
We'll send you Level 2 instructions as well.
Send us email (pointerwars2025@gmail.com) and we will send along level 2 instructions for you. If you have changed your mind and want code and performance numbers, you are more than welcome to send mail and we will send you an invoice to upgrade to the paid tier.
Some general advice on things to watch out for in Level 2 will be provided to you as well.
Level 2 will provide you with software you can run locally to test how fast your code runs. With a few tweaks and best practices, you can use this software to determine what a change does in terms of performance.
Paid participants have their runs performed on our machines, so they can be compared against others, and also receive PMU counter feedback, along with interpretation of those PMU counters, in level 2.