compiling functions....

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

compiling functions....

akshay kulkarni
dear members,
                            I a Day Trader based in INDIA. I use R for my research. I have a function ygusa(snlq,snlcqn) which takes 208 stocks and returns 4 best stocks for the next day(snlq is the list of 208 stocks and snlcqn is their names). However, the execution time is around 2 hrs, making it hard for me.

I recently read in the internet that you can precompile the code in R to make it run faster. Also that you can enable JIT(just in time compilation) from your R session automatically. I came to know that R 3.4.x has JIT enabled in it by default. Is it true? Is it also true that even after enabling JIT in R 3.4.x, the first run of a function is not Byte compiled?

So when I start my R session, download the data, and run ygusa(snlq,snlcqn), it is not byte compiled and therefore it is very slow? Will including the following lines in ygusa solve my problem:
> require(compiler)
> enableJIT(3)

?

Also, instead of compiling the function ygusa every time I run it, can I compile it once and store it, and run that compiled file instead of ygusa(snlq,snlcqn)?

Can you point me to some online resources that can help on this issue?

very many thanks for your time and effort...
Yours sincerely,
AKSHAY M KULKARNI

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: compiling functions....

Sarah Goslee
There are many things you can do to improve speed in R. Byte compiling
is just one of them.

This chapter in Hadley Wickham's excellent Advanced R book covers both
profiling and byte compiling.

http://adv-r.had.co.nz/Profiling.html

I've gotten some stunning improvements in speed through profiling and
careful thought: 3 days to 3 seconds, even.

Sarah

On Thu, Jun 21, 2018 at 6:28 AM, akshay kulkarni <[hidden email]> wrote:

> dear members,
>                             I a Day Trader based in INDIA. I use R for my research. I have a function ygusa(snlq,snlcqn) which takes 208 stocks and returns 4 best stocks for the next day(snlq is the list of 208 stocks and snlcqn is their names). However, the execution time is around 2 hrs, making it hard for me.
>
> I recently read in the internet that you can precompile the code in R to make it run faster. Also that you can enable JIT(just in time compilation) from your R session automatically. I came to know that R 3.4.x has JIT enabled in it by default. Is it true? Is it also true that even after enabling JIT in R 3.4.x, the first run of a function is not Byte compiled?
>
> So when I start my R session, download the data, and run ygusa(snlq,snlcqn), it is not byte compiled and therefore it is very slow? Will including the following lines in ygusa solve my problem:
>> require(compiler)
>> enableJIT(3)
>
> ?
>
> Also, instead of compiling the function ygusa every time I run it, can I compile it once and store it, and run that compiled file instead of ygusa(snlq,snlcqn)?
>
> Can you point me to some online resources that can help on this issue?
>
> very many thanks for your time and effort...
> Yours sincerely,
> AKSHAY M KULKARNI
>


--
Sarah Goslee
http://www.functionaldiversity.org

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.