Delicious Timer is a fast, lightweight profiler oriented towards execution speed.

It makes it easy to visualise what happens in your program and lets you quickly spot opportunities for speed-ups.

Some key points:

Delicious Timer is fast: it can record up to millions of time lapses per second.
It is easy to navigate: it feels snappy and responsive, even when there is tons of data to display.
You can get the statistics you care about: the visualiser lets you instantly combine threads, compute local statistics, sort results and search through them. You can also highlight locks and memory allocations throughout your program.
Get it up and running within minutes: the Delicious Timer API is simple and straightforward. You can disable it at any time with one line of code.
Easily record time lapses, context switches, thread locks and memory allocations.
Bloat-free: Delicious Timer is fully portable and weighs less than a megabyte.

Requirements:

For now only Windows is supported, there are plans for a Linux version.

Windows 10, 11 (other versions remain untested)
A modern AMD64 (x64) CPU (anything from the last 10 years should do)
Lots of RAM if you plan to record tons of time lapses (memory usage is proportional to the number of time lapses recorded)

v1.2.0 release notes:

There are now C# and Rust ports.

Some changes were made to the API to make it more usable.

Fixed a bug where you could not close the profiler from the task bar.

Some parts of the profiler were rewritten from scratch for this release to simplify systems and get rid of bugs.