coz VS stabilizer

But let's take the guesswork out of profilers. Use Coz. It's a causal profiler that performs experiments to determine what code would see the greatest performance improvement of the whole program if made faster. There's a video in the link; I think the best demonstration they had was a program that saw the greatest improvement by optimizing a function that ranked #30 by a sampling profiler.

Coz. It's in a Debian package so you don't have to build it. Watch the video embedded in the page I linked; I;m all about profiling, but the devil is if you're not a statistician, you don't know how to read profiler results.

I think https://github.com/plasma-umass/coz solves most problems related to noise in benchmarks.

Perhaps you're thinking of the coz profiler (https://github.com/plasma-umass/coz)?

I was thinking of doing some kernel profiling, and stumbled upon this interesting repo: https://github.com/plasma-umass/coz

I'm pretty intrigued by the concept, and was wondering if anyone here tried out Coz.

https://github.com/plasma-umass/coz https://youtu.be/7g1Acy5eGbE

Could it be an issue of binary layout? Have a look at the coz profiler which has a rust port.

Yes, just because it's from a team connected with the old Stabilizer tool, it doesn't mean it shares the insights and benefits of that kind of testing. At least, that's my understanding of coz, based on my knowledge of https://github.com/ccurtsinger/stabilizer

Nope, he is thinking of that team's previous work, stabilizer. See this talk for more information on both. It is an excellent talk.

Other researchers have tried building tools to address this issue : stabilizer. Unfortunately, stabilizer is now outdated as it hasn't been upgraded from it's LLVM 3.1 foundation. See the project here : https://github.com/ccurtsinger/stabilizer

I see a lot a value in stabilizer.

Any thoughts on using stabilizer in the future?: https://github.com/ccurtsinger/stabilizer