Ece 459 Lecture 28 Cachegrind

The last of the Valgrind tools we will cover is the sadly non-Norse-named Causal profiling allows for running a what-if kind of assessment to understand the impact -- positive, negative, or none at all -- of ... We'll learn about Massif, the heap profiler in the Valgrind toolset, and its newer alternative, DHAT. In our last hardware-focused topic, we'll consider how CPUs that share main memory but have individual caches can all manage ... Are you tired of grinding leetcode? I mean, who isn't? Here, we'll discuss a bit about algorithmic complexity. This isn't about how ... We'll learn about Valgrind, which although is much more intended for languages like C/C++, can still be put to use in Rust in ...

Finally, we're ready to talk about profiling tools -- identifying what's slow in the program. We can use the profiler to help the compiler make good decisions when compiling; let's find out how! We will start our discussion of profiling with tracing, which is to say logging, or maybe just printf with extra steps. Continuing our hardware discussion, we'll talk about cache misses and why they're so expensive in terms of performance. This is a look into how parallelism really helps us get things done, the limits of it, and how we might move beyond them. This is a brief introduction to the course and a tiny little bit about me.

This is an introduction to the course aas well as putting some definitions to the term "performance" and giving a high-level ...