I was curious how much better the Algorithms, pointed out by BitTwiddling Hacks, perform than a simple Lookup Table approach...
Now I think, perhaps my little investigation is interesting for others too...
PS: the parallel counting Algo is about 35% faster on average on my computers than a simple LUT based solution. Also this shows nicely how different a human brain compatible solution is to a binary machine optimal solution :-)
In the meantime I added other Algorithms beside countbisset (hamming weight): ceil(log2()) + floor(log2()), bitreverse, ...
I also recommand to try the code with different compilers and / or plattforms.
PS: see result.txt in Code, to see various runs with different compilers and compiler switches
Features
- C++
- CountBitsSet, PopulationCount.Integer Hamming weight
- simple Benchmark
Categories
AlgorithmsLicense
Public DomainFollow CountBitsSet
Other Useful Business Software
Get the most trusted enterprise browser
Defend against security incidents with Chrome Enterprise. Create customizable controls, manage extensions and set proactive alerts to keep your data and employees protected without slowing down productivity.
Rate This Project
Login To Rate This Project
User Reviews
Be the first to post a review of CountBitsSet!