Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 7 years ago.
Improve this question
There are 2 ways to run the R parallelly, which are SparkR and RevolutionR.
From the compatibility with R,distribution processing effect, scalability, application scenarios, community support and maybe some other aspects, what are the real difference between these two?
As I known, the biggest difference is that Revolution R is a commercial software while SparkR is free. So, you even can't try most of the parallel functions of Revolution R before paid.
Wait for other guys w/ real experience of Revolution R to update the answer :)
Btw, there are not only 2 ways to run R parallel. Other approaches, such as snow, multicore, parallel, Rmpi, ... And offload to GPU as here.
Check out below two links:
RevolutionR
SparkR
Related
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 3 years ago.
Improve this question
I am not able to find any definitive answer to what problem Julia solves compared to the two languages, I was told are most commonly used in working with data (data science), R and Python.
I am not asking for any opinions. Please support replies with factual information (sad I need to add this but it seems some people think this topic is opinion based).
Could anyone explain this?
The Julia Programming Language solves the same problems as R and Python. However, it can solve them extremely faster than those mentioned above, as it runs over C code and uses a JIT compiler. See the Julia Benchmark. This and other advantages that can be found at the language site, it's Twitter profiles: Julia Computing and Julia Language.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 6 years ago.
Improve this question
I have looked around and have not found many opinions on if it is important to update your R software as soon as a new version is released.
Any opinions would be welcomed!
As with any software, you should carefully evaluate what is included in any new release. If the release consists only of bug-fixes, it is usually expedient to install it as soon as it is practicable for you to do so. If the scope of the release is more expansive -- new features, etc. -- you should review the release more carefully.
If you're in the middle of an important project with a killer deadline, it's quite reasonable to wait a little while before applying any update.
Also, you should as a matter of routine re-run a selection of jobs, that you know the answers to, in order to be sure that the answers are still the same. "No, mistakes of this nature don't happen often, but they do happen."
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 5 years ago.
Improve this question
I know this is not precisely a programming question, but I don't where else to ask...
S-PLUS was aquired by TIBCO some years ago. And it was seemingly included to the Spotfire product. However I installed the demo version of Spotfire and can't find anything indicating that S-PLUS is anymore part of it.
So my question is: is S-PLUS dead? And is there any way to install a prior version of it? I know R has totally taken over, but I'd be curious to just try it out if it is available somewhere.
Not sure how relevant the question is but here are my $0.02:
Yes, R has won.
TIBCO still seems to have a Spotfire product mentioning S-Plus (pdf found via simple Google search).
IIRC, years ago TIBCO purchased the commercial S license, but it turns out that nobody really wanted S-Plus if it was not entirely compatible with R.
TIBCO learned that lesson and built an entirely new R-compatible engine they call TERR; the jury is still out as to whether it will ever get any significant traction.
In the end, it is rather difficult to beat a well-maintained and written Open Source product---and R is one of the better examples of Open Source done right.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 9 years ago.
Improve this question
The question is in the Title. Basically I'm looking for an alternative to R.
I've been using R a bit, there are some really good stuff about it (especially data.frame plyr and ggplot), however I really love Haskell and type inference, so I was wondering if using Haskell to do "simple" statistic analysis would be a good choice.
My basic needs are :
read/write CSV
import SQL table
do some basic 'mapReduce' on the data. Which where R is great but I assume Haskell should be equally good.
However my experience with Haskell is everything is fine until you process realworld data. You always encounter performance issue (and soonish) because even though in theory you should write functional code and don't worry about what's the computer is doing, if you don't and don't use the appropriate library and are not an Haskell expert, stuff are damned slow.
Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 10 years ago.
Improve this question
I know R, I know SQL, I use Windows, I have a budget of $0, I have a terabyte of data, I have twelve processors, I have 96GB of RAM, I am motivated to learn new software if the speed gains will pay off in the long term.
I need to run descriptive statistics and regressions.
I have too many options. Where should I devote all of my energy? Thanks.
Well, that is a big topic.
We did write a survey paper of the state of the art of parallel processing with R which you could start with. While it is now three years old, large parts of the discussion still hold.
Otherwise, I would suggest starting small- to medium-size with something that actually matters to you and try to get that going faster. Over at the r-sig-HPC list (gmane link) list many folks are happy to help with specifics.