Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

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

Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Juan Telleria
Dear R Developers,

First of all, I would like to thank you Jeroen Ooms for taking the binary
Window Builds from Duncan. I firmly believe that the R Community will
benefit a lot from his work.

However, the debate I would like to open is about if some of Microsoft R
Open Code shall be ported from R Open to Mainstream R.

There are some beneficts in R Open such as multithreaded performance:
https://mran.microsoft.com/documents/rro/multithread/

Maybe, the R Consortium, and in particular, Microsoft R Team, could
collaborate, if appropriate, in such duty.

Thank you,
Juan Telleria

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Kenny Bell-2
User here: incorporating Intel's MKL, as MRO does, would be a very welcome
addition.

I was an MRO user before and it improved my experience with medium data
immensely.

They did, however, leave behind bugs here and there, especially related to
development with Rcpp, so I switched back to vanilla R.

On Mon, Oct 30, 2017, 9:42 AM Juan Telleria <[hidden email]> wrote:

> Dear R Developers,
>
> First of all, I would like to thank you Jeroen Ooms for taking the binary
> Window Builds from Duncan. I firmly believe that the R Community will
> benefit a lot from his work.
>
> However, the debate I would like to open is about if some of Microsoft R
> Open Code shall be ported from R Open to Mainstream R.
>
> There are some beneficts in R Open such as multithreaded performance:
> https://mran.microsoft.com/documents/rro/multithread/
>
> Maybe, the R Consortium, and in particular, Microsoft R Team, could
> collaborate, if appropriate, in such duty.
>
> Thank you,
> Juan Telleria
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Dirk Eddelbuettel

On 29 October 2017 at 22:01, Kenny Bell wrote:
| User here: incorporating Intel's MKL, as MRO does, would be a very welcome
| addition.
|
| I was an MRO user before and it improved my experience with medium data
| immensely.
|
| They did, however, leave behind bugs here and there, especially related to
| development with Rcpp, so I switched back to vanilla R.

With all due respect: You may miss something. MKL has always worked with 'Base R'.

As a point of reference and comparison, I set up a benchmarking and
comparison package _well over half a decade ago_ and while it never get fully
finished to the point of a submitted paper the vignette still stands---and
demonstrates that _dropping in MKL is a one-line operation_.

And always has been.  There may have been some license arbitrage: Intel was
an early investor in Revo, so MKL was pushed hard.  With GotoBLAS and later
OpenBLAS I cared less, but IIRC the license of MKL is a little simpler for
"mere use" now.

See  https://cloud.r-project.org/web/packages/gcbd/vignettes/gcbd.pdf  for more.

Hth,  Dirk

|
| On Mon, Oct 30, 2017, 9:42 AM Juan Telleria <[hidden email]> wrote:
|
| > Dear R Developers,
| >
| > First of all, I would like to thank you Jeroen Ooms for taking the binary
| > Window Builds from Duncan. I firmly believe that the R Community will
| > benefit a lot from his work.
| >
| > However, the debate I would like to open is about if some of Microsoft R
| > Open Code shall be ported from R Open to Mainstream R.
| >
| > There are some beneficts in R Open such as multithreaded performance:
| > https://mran.microsoft.com/documents/rro/multithread/
| >
| > Maybe, the R Consortium, and in particular, Microsoft R Team, could
| > collaborate, if appropriate, in such duty.
| >
| > Thank you,
| > Juan Telleria
| >
| >         [[alternative HTML version deleted]]
| >
| > ______________________________________________
| > [hidden email] mailing list
| > https://stat.ethz.ch/mailman/listinfo/r-devel
| >
|
| [[alternative HTML version deleted]]
|
| ______________________________________________
| [hidden email] mailing list
| https://stat.ethz.ch/mailman/listinfo/r-devel

--
http://dirk.eddelbuettel.com | @eddelbuettel | [hidden email]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Ista Zahn
On Sun, Oct 29, 2017 at 11:09 PM, Dirk Eddelbuettel <[hidden email]> wrote:

>
> On 29 October 2017 at 22:01, Kenny Bell wrote:
> | User here: incorporating Intel's MKL, as MRO does, would be a very welcome
> | addition.
> |
> | I was an MRO user before and it improved my experience with medium data
> | immensely.
> |
> | They did, however, leave behind bugs here and there, especially related to
> | development with Rcpp, so I switched back to vanilla R.
>
> With all due respect: You may miss something. MKL has always worked with 'Base R'.
>
> As a point of reference and comparison, I set up a benchmarking and
> comparison package _well over half a decade ago_ and while it never get fully
> finished to the point of a submitted paper the vignette still stands---and
> demonstrates that _dropping in MKL is a one-line operation_.
>
> And always has been.  There may have been some license arbitrage: Intel was
> an early investor in Revo, so MKL was pushed hard.  With GotoBLAS and later
> OpenBLAS I cared less, but IIRC the license of MKL is a little simpler for
> "mere use" now.
>
> See  https://cloud.r-project.org/web/packages/gcbd/vignettes/gcbd.pdf  for more.

FWIW, I've been using openBlas for years now, based on this and other
benchmarks. It provides performance comparable to MKL while being
really free.

Best,
Ista

>
> Hth,  Dirk
>
> |
> | On Mon, Oct 30, 2017, 9:42 AM Juan Telleria <[hidden email]> wrote:
> |
> | > Dear R Developers,
> | >
> | > First of all, I would like to thank you Jeroen Ooms for taking the binary
> | > Window Builds from Duncan. I firmly believe that the R Community will
> | > benefit a lot from his work.
> | >
> | > However, the debate I would like to open is about if some of Microsoft R
> | > Open Code shall be ported from R Open to Mainstream R.
> | >
> | > There are some beneficts in R Open such as multithreaded performance:
> | > https://mran.microsoft.com/documents/rro/multithread/
> | >
> | > Maybe, the R Consortium, and in particular, Microsoft R Team, could
> | > collaborate, if appropriate, in such duty.
> | >
> | > Thank you,
> | > Juan Telleria
> | >
> | >         [[alternative HTML version deleted]]
> | >
> | > ______________________________________________
> | > [hidden email] mailing list
> | > https://stat.ethz.ch/mailman/listinfo/r-devel
> | >
> |
> |       [[alternative HTML version deleted]]
> |
> | ______________________________________________
> | [hidden email] mailing list
> | https://stat.ethz.ch/mailman/listinfo/r-devel
>
> --
> http://dirk.eddelbuettel.com | @eddelbuettel | [hidden email]
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Avraham Adler
In reply to this post by Kenny Bell-2
[Sent offlist accidentally]

What concerns me first and foremost is that the licensure would have
to be ironclad (including for commercial use like vanilla R now) as
well as ensuring that R remains completely FLOSS. Anything “added” to
R has to be a no-strings-attached gift to R.

Also, I would think that it would have to play nice with existing
workflows (like OpenBLAS instead of MKL) unless there is such a
benefit that it is worth breaking compatibility.

Avi

On Sun, Oct 29, 2017 at 6:01 PM, Kenny Bell <[hidden email]> wrote:

> User here: incorporating Intel's MKL, as MRO does, would be a very welcome
> addition.
>
> I was an MRO user before and it improved my experience with medium data
> immensely.
>
> They did, however, leave behind bugs here and there, especially related to
> development with Rcpp, so I switched back to vanilla R.
>
> On Mon, Oct 30, 2017, 9:42 AM Juan Telleria <[hidden email]> wrote:
>
>> Dear R Developers,
>>
>> First of all, I would like to thank you Jeroen Ooms for taking the binary
>> Window Builds from Duncan. I firmly believe that the R Community will
>> benefit a lot from his work.
>>
>> However, the debate I would like to open is about if some of Microsoft R
>> Open Code shall be ported from R Open to Mainstream R.
>>
>> There are some beneficts in R Open such as multithreaded performance:
>> https://mran.microsoft.com/documents/rro/multithread/
>>
>> Maybe, the R Consortium, and in particular, Microsoft R Team, could
>> collaborate, if appropriate, in such duty.
>>
>> Thank you,
>> Juan Telleria
>>
>>         [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> [hidden email] mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Juan Telleria
So as long as I can read, OpenBlas, for Windows, might be a worth
considering option:
http://www.openblas.net

But Intel MKL also seems to be free*:
https://software.intel.com/en-us/articles/free-mkl

Thank you,
Juan

El 30/10/2017 6:45 p. m., "Avraham Adler" <[hidden email]>
escribió:

> [Sent offlist accidentally]
>
> What concerns me first and foremost is that the licensure would have
> to be ironclad (including for commercial use like vanilla R now) as
> well as ensuring that R remains completely FLOSS. Anything “added” to
> R has to be a no-strings-attached gift to R.
>
> Also, I would think that it would have to play nice with existing
> workflows (like OpenBLAS instead of MKL) unless there is such a
> benefit that it is worth breaking compatibility.
>
> Avi
>
> On Sun, Oct 29, 2017 at 6:01 PM, Kenny Bell <[hidden email]> wrote:
> > User here: incorporating Intel's MKL, as MRO does, would be a very
> welcome
> > addition.
> >
> > I was an MRO user before and it improved my experience with medium data
> > immensely.
> >
> > They did, however, leave behind bugs here and there, especially related
> to
> > development with Rcpp, so I switched back to vanilla R.
> >
> > On Mon, Oct 30, 2017, 9:42 AM Juan Telleria <[hidden email]>
> wrote:
> >
> >> Dear R Developers,
> >>
> >> First of all, I would like to thank you Jeroen Ooms for taking the
> binary
> >> Window Builds from Duncan. I firmly believe that the R Community will
> >> benefit a lot from his work.
> >>
> >> However, the debate I would like to open is about if some of Microsoft R
> >> Open Code shall be ported from R Open to Mainstream R.
> >>
> >> There are some beneficts in R Open such as multithreaded performance:
> >> https://mran.microsoft.com/documents/rro/multithread/
> >>
> >> Maybe, the R Consortium, and in particular, Microsoft R Team, could
> >> collaborate, if appropriate, in such duty.
> >>
> >> Thank you,
> >> Juan Telleria
> >>
> >>         [[alternative HTML version deleted]]
> >>
> >> ______________________________________________
> >> [hidden email] mailing list
> >> https://stat.ethz.ch/mailman/listinfo/r-devel
> >>
> >
> >         [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > [hidden email] mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-devel
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Iñaki Úcar
2017-10-31 14:34 GMT+01:00 Juan Telleria <[hidden email]>:
> So as long as I can read, OpenBlas, for Windows, might be a worth
> considering option:
> http://www.openblas.net
>
> But Intel MKL also seems to be free*:
> https://software.intel.com/en-us/articles/free-mkl

install.packages("rmsfact")
sub(".*because ", "", rmsfact::rmsfact(8))

Iñaki

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Juan Telleria
OpenBlas seems to have a performance similar to Intel MKL, as is stated in
Wikipedia:

https://en.m.wikipedia.org/wiki/OpenBLAS

However, I would suggest debating this topic in the R Consortium and the R
Foundation, for taking the best decision possible for the future or R.

Thank you all,
Juan

El 31/10/2017 2:55 p. m., "Iñaki Úcar" <[hidden email]> escribió:

2017-10-31 14:34 GMT+01:00 Juan Telleria <[hidden email]>:
> So as long as I can read, OpenBlas, for Windows, might be a worth
> considering option:
> http://www.openblas.net
>
> But Intel MKL also seems to be free*:
> https://software.intel.com/en-us/articles/free-mkl

install.packages("rmsfact")
sub(".*because ", "", rmsfact::rmsfact(8))

Iñaki

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: Debate: Shall some of Microsoft R Open Code be ported to mainstream R?

Martin Maechler
In reply to this post by Iñaki Úcar
>>>>> Iñaki Úcar <[hidden email]>
>>>>>     on Tue, 31 Oct 2017 14:55:44 +0100 writes:

    > 2017-10-31 14:34 GMT+01:00 Juan Telleria
    > <[hidden email]>:
    >> So as long as I can read, OpenBlas, for Windows, might be
    >> a worth considering option: http://www.openblas.net
    >>
    >> But Intel MKL also seems to be free*:
    >> https://software.intel.com/en-us/articles/free-mkl

    > install.packages("rmsfact")
    > sub(".*because ", "", rmsfact::rmsfact(8))

"Amen"!

... and thank you Iñaki  for alerting us to the rmsfact package.
Cool!  

Martin Maechler
ETH Zurich and R Core Team

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel