Pavel Samolysov

@avijit-nervana Could you have a look at the PR, please? Maybe I wasn't able to properly titled the PR, but it is very important because without this change any `DataType`s...

I have no access to the CI server and unfortunately am not able to see what is wrong with the commits.

@avijit-nervana I've updates the comment. Could you please merge the PR if there is no concerns more? I'm working on OpenSUSE and making manual changing after each `git pull` makes...

@avijit-nervana What can we do to merge the PR? Do I have to sign any contribution agreement or something like this?

I agree with your answer and have proposed a PR #295. If we have a look at fig 18.1, we have a chance to see something a bit similar: ```...

As I see, this branch has not been rebased on top of the recently merged changes with the introduction of the `createDummyReturn` function (or the final version of its name)....

@jeremy-murphy Thank you very much for the answer and a lot of information and suggestion. I believe developing benchmarks with the google benchmark project ismay itself be a good first...

@jeremy-murphy Hello, sorry for a late response. I would like to implement this point: >> Changing the type of buckets_ to boost::unordered_flat_map, and instead of clearing the bucket, remove it...

@jeremy-murphy About this suggestion: > Changing the type of buckets_ to boost::unordered_flat_map, and instead of clearing the bucket, remove it from the hash map. I've replaced `std::vector< std::vector > buckets_`...

I use numerous small test cases (10-14 vertices) from the `dominator_tree_test.cpp` test as well as a 10x times larger example - a CFG from a real code (up to 200...