flamegraph VS cargo-asm

You can install cargo-flamegraph with cargo install flamegraph. There are some underlying requirements to be able to use cargo-flamegraph; you will want to take a look at the repo here to make sure you have the right dependencies.

I tried to diagnose the issue with flamegraph, but unfortunately the flamegraph didn't show anything beyond the next call for some reason

having some tools to identify slowness origins (flamegraph is one... but not sure it's the way to go)

I'm currently implementing a solution for the first part of the day 16. It work but it is really slow... I'd like to : - understand why - having some tools to identify slowness origins (flamegraph is one... but not sure it's the way to go) - eventually have some clue/solution/idea - have general feedback on what in my "coding style" is not appropriate for rust (I come from java/kotlin/ts even if I've already coded a bit in c/c++) : for example I love iterator & sequence but i feel they are not really suited to overuse in rust (mostly because of async & result).

Use a profiler. Flamegraph is a good way to visualise profiler output. This lets you identify which functions are taking up a large amount of time - and hence helps you identify where to focus your optimisation efforts.

You tried trying to see what takes the most time under load via flames? https://github.com/flamegraph-rs/flamegraph

Well, in cases like that the answer is straight forward, use a profiler like https://github.com/flamegraph-rs/flamegraph

I don't see any obvious problems with your algorithm. I've had luck using cargo-flamegraph to identify the slow parts of my code. That's going to show you which parts to focus on improving the performance of!

From the readme of cargo flamegraph:

cargo asm might be useful here (if you can't use godbolt).

They should be inlined in release mode. You'd have to verify by checking the assembly, though, which could be done directly in the Rust playground or with a tool like cargo-asm.

You can use cargo asm - not sure if you can integrate it with VSCode, but even from a terminal it's a pretty convenient tool.

There's probably a built-in one somewhere, but I suspect it'd be easier just to install https://github.com/gnzlbg/cargo-asm

I don't know if there's a way to do a side by side comparison but cargo-asm uses the source mapping information from rustc to annotate chunks of assembly with it's respective rust code, though it's an imperfect process.

After that you would need some tools to help figure out how to achieve improvements. That will depend on your system and personal preferences. As the other commenter suggested, perf is a good choice on linux. I personally like to look at the generated assembly, using either cargo asm, godbolt, or just rust playground.

Doesn't this already exist? https://github.com/gnzlbg/cargo-asm

From the flamegraph, we knew which function was taking the majority of the time, which we briefly described above. We started by looking at the assembly code it generated to see if there were any opportunities to make it faster. We did this with cargo-asm.

You're best bet at the moment is probably using cargo-asm to inspect the function assembly to see when it is performing the correct TCO.