@John_Gibson For Stata 14 and Stata 15 on OpenSuse, there is a Stata bug and preliminary fix related to libpng. This will be addressed officially in a Stata update.
Actually the issue is the license. My university (U New Hampshire) does not have a site license for Stata, and I’m not inclined to purchase one just to include it in the benchmarks.
1 Like
You can try Stata for free for one month, enough time to include it one time in the benchmarks:
I am really curious how Stata compares with R and Python in the benchmarks.
Including Stata one time is of fairly limited value and IMO not worth the effort. Our conclusion the last time this was discussed is that there was no reason to believe that Stata would have particularly interesting performance characteristics that would differ much compared to other slow dynamic language implementations. My very uneducated guess is that it would be more similar to R (one of the slower interpreted languages) than Python (one of the faster ones). If you want to compare Stata with R and Python, you can run benchmarks for those three languages on your own hardware.
Despite where people seem to want to take this, we’re not in the business of evaluating the performance of all possible languages. We really only care about showing that Julia is as fast as the fastest languages out there and that it is much faster at basic programming techniques like recursion, iteration, etc. than languages like Matlab, R and Python that are commonly used in numerical computing. That point has been pretty well established by now. It’s not terribly useful or interesting to add yet another standard interpreter implementation to the list.
9 Likes
@StefanKarpinski Fair enough. Matthieu Gomez compared the speed of R and Stata with a different set of benchmarks and, as Stefan guessed, the overall results are similar:
Hi all,
I noticed this 2 years old study “Energy Efficiency across Programming Languages How Does Energy, Time, and Memory Relate?”
It could be interesting to know how Julia performs in terms of energy consumption.
Kind regards
1 Like
That’s a really interesting document. I fear that in case of Julia it will depend… If you “rerun” the code several times, you will have to compile a lot over and over again (not sure about the status of the precompilation cache though), which means that Julia cannot be compared to statically compiled languages per se, but correct me if I am wrong. With an efficient (pre)compilation cache, it will be comparable I think.