Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
Folks: 1. No blame implied to anyone. However, as 64 bit Windows/R build is apparently not available except as a commercial product, may I suggest that henceforth it should not be discussed in this list and that any queries about it simply be directed to David Smith at Revolution. It just doesn't "feel" right to me to have these sorts of discussions here. I know that no commercialism was intended, but it still seems to me be oozing in. Bert Gunter Genentech Nonclinical Biostatistics -Original Message- From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On Behalf Of David M Smith Sent: Tuesday, October 06, 2009 11:57 AM To: Jose Quesada Cc: r-h...@stat.math.ethz.ch Subject: Re: [R] R on Linux, and R on Windows ,any difference in maturity+stability? I wanted to correct a couple of misconceptions raised in Jose's post below, which I'll take the liberty of addressing out of order. On Tue, Oct 6, 2009 at 7:34 AM, Jose Quesada wrote: > While revolution has > provided very nice packages to the community (e.g., foreach), the win-64 port as > of today is certainly the worst platform to do work on. Reasons: > (1) it's R 2.7.2 That's true, however REvolution R Enterprise based on R 2.9.2 is in beta testing right now. Creating a distribution of R under validated build processes adds a lot of extra process, testing, and overhead, and for this and other reasons our subscription "Enterprise" distributions don't track R version-to-version. Our free, community-based REvolution R distributions will track R much more closely beginning with 2.9.2. > (4) There's a proprietary repository, where most packages are outrageously > outdated. As a service to our users using REvolution R Enterprise we provide 64-bit Windows binary builds of as many R packages on CRAN as possible (see below). This is helpful to many of our users, because the tools to build binary packages for R on 64-bit Windows are not widely available. (Unfortunately, no free compiler is capable of building R for 64-bit Windows today. Believe me, we and many others have tried.) CRAN does not support Windows 64-bit binary packages, so we must provide a repository separate from CRAN (and our own CRAN mirror) for this purpose. But calling it a "proprietary repository" misleads -- all those packages are and remain free under the terms of their respective licenses (GPL and others). The packages are all compatible with R 2.7.2, which is currently the only 64-bit Windows version of R available. > (2) Many important packages will never be ported Some clarification is in order here. A number of packages on CRAN are not self-contained; some rely on third-party software or systems not part of R itself. For example, RGtk2 depends on the gtk+ software, which is only available in experimental form on the Windows 64-bit platform. This, obviously, has ramifications for the packages that they depend on. In some cases our support team has gone above and beyond for subscription customers to port third-party applications (for example, we ported the independent GraphViz software to 64-bit Windows to make RGraphViz from BioConductor work), but for obvious reasons this can only be done on a case-by-case basis. > (3) Some packages (particularly those depending on Rjava) would not work properly. See (2) above: as a contributed package, rJava is dependent on Windows' support for Java on the 64-bit platform. Some have noted that Microsoft's love for Java is less than legendary. > (5) Most help you find on R-help will not apply. Instead, you have 'paid' > support. Said support is slow, and close to useless in most cases. Jose is entitled to his opinion, but the live technical support provided by our team is a major feature of our subscription-based distributions - it is indeed what you pay for. We have many customers from commercial institutions large and small, on Windows 64-bit and other platforms, who have found great value in the responsiveness and expertise of our support services. As I've noted above, in many cases they go beyond the call of duty to deal with issues inherent to the Windows 64-bit operating system. Unfortunately, platform-specific issues are sometimes beyond our control, despite best efforts. > (6) Packages that rely on external tools (e.g., mysql) will take a lot of work > to get going. This is true of many software packages for 64-bit Windows including mysql. Unfortunately, the lack of good free compilers for the Windows 64-bit platform means that some open-source projects in particular are not readily available for 64-bit Windows (a situation we at REvolution Computing seek to remedy for R). > And of course, one have to pay for a yearly license, to have the privilege to > work under the above conditions. As Heinlein wrote, TANSTAAFL. The feedback we've had is that support for
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
I wanted to correct a couple of misconceptions raised in Jose's post below, which I'll take the liberty of addressing out of order. On Tue, Oct 6, 2009 at 7:34 AM, Jose Quesada wrote: > While revolution has > provided very nice packages to the community (e.g., foreach), the win-64 port > as > of today is certainly the worst platform to do work on. Reasons: > (1) it's R 2.7.2 That's true, however REvolution R Enterprise based on R 2.9.2 is in beta testing right now. Creating a distribution of R under validated build processes adds a lot of extra process, testing, and overhead, and for this and other reasons our subscription "Enterprise" distributions don't track R version-to-version. Our free, community-based REvolution R distributions will track R much more closely beginning with 2.9.2. > (4) There's a proprietary repository, where most packages are outrageously > outdated. As a service to our users using REvolution R Enterprise we provide 64-bit Windows binary builds of as many R packages on CRAN as possible (see below). This is helpful to many of our users, because the tools to build binary packages for R on 64-bit Windows are not widely available. (Unfortunately, no free compiler is capable of building R for 64-bit Windows today. Believe me, we and many others have tried.) CRAN does not support Windows 64-bit binary packages, so we must provide a repository separate from CRAN (and our own CRAN mirror) for this purpose. But calling it a "proprietary repository" misleads -- all those packages are and remain free under the terms of their respective licenses (GPL and others). The packages are all compatible with R 2.7.2, which is currently the only 64-bit Windows version of R available. > (2) Many important packages will never be ported Some clarification is in order here. A number of packages on CRAN are not self-contained; some rely on third-party software or systems not part of R itself. For example, RGtk2 depends on the gtk+ software, which is only available in experimental form on the Windows 64-bit platform. This, obviously, has ramifications for the packages that they depend on. In some cases our support team has gone above and beyond for subscription customers to port third-party applications (for example, we ported the independent GraphViz software to 64-bit Windows to make RGraphViz from BioConductor work), but for obvious reasons this can only be done on a case-by-case basis. > (3) Some packages (particularly those depending on Rjava) would not work > properly. See (2) above: as a contributed package, rJava is dependent on Windows' support for Java on the 64-bit platform. Some have noted that Microsoft's love for Java is less than legendary. > (5) Most help you find on R-help will not apply. Instead, you have 'paid' > support. Said support is slow, and close to useless in most cases. Jose is entitled to his opinion, but the live technical support provided by our team is a major feature of our subscription-based distributions - it is indeed what you pay for. We have many customers from commercial institutions large and small, on Windows 64-bit and other platforms, who have found great value in the responsiveness and expertise of our support services. As I've noted above, in many cases they go beyond the call of duty to deal with issues inherent to the Windows 64-bit operating system. Unfortunately, platform-specific issues are sometimes beyond our control, despite best efforts. > (6) Packages that rely on external tools (e.g., mysql) will take a lot of work > to get going. This is true of many software packages for 64-bit Windows including mysql. Unfortunately, the lack of good free compilers for the Windows 64-bit platform means that some open-source projects in particular are not readily available for 64-bit Windows (a situation we at REvolution Computing seek to remedy for R). > And of course, one have to pay for a yearly license, to have the privilege to > work under the above conditions. As Heinlein wrote, TANSTAAFL. The feedback we've had is that support for R is particularly beneficial on the 64-bit Windows platform, for exactly the reasons Jose raises above. > Note: this may change any time, since they are working on a continuous build > that will keep the releases in sync with mainstream R. That's also true. We have been working on an automated build system for R and R packages. It will first be used for the Ubuntu release, but it's designed as a multiplatform system. It will allow us to keep our free community releases in sync with CRAN R, and keep packages we build up-to-date. In summary: I thank Jose for raising some important points to consider if you have the freedom to choose any platform and you need the additional memory capacity of a 64-bit version of R. But if you need to use 64-bit Windows and want to use R, the supported distribution REvolution R Enterprise has worked very well for many users. # David Smith -- David M Smith Director of Community, REvo
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
On 10/6/2009 1:43 PM, David M Smith wrote: On Tue, Oct 6, 2009 at 10:30 AM, Duncan Murdoch wrote: When running a 32 bit program, 64 bit Windows hides most of itself outside the address space visible to the program, so almost all of the 4 Gb address space is available to the user. But no more: no matter how much RAM you install, it's not possible to address it using a 32 bit address. That's true when running a 32-bit version of R under 64-bit Windows. But on the 64-bit version of REvolution R Enterprise for Windows, you can create objects much larger than 4Gb, for example: memory.limit(2e+9) v<-double(1e+9) # 8GB vector sum(v) [1] 0 The same is true of 64-bit versions of R on other platforms, too. Yes, I was only talking about running 32 bit R. 64 bit R has another advantage on all platforms: since we're still using such a small fraction of the whole 64 bit address space, fragmentation isn't such a problem as it is with 32 bits. For those unfamiliar with the term: fragmentation happens when you allocate things at particular addresses, limiting the size of future allocations to the space available above or below or between those allocations. For the allocation David did, R needs a full 8Gb of memory at contiguous addresses. The hardware can remap physical or virtual memory to any 64 bit address it likes (within some fairly relaxed limits), so that's not a problem in 64 bit R. In 32 bit R, it means that a few small objects could easily cut your maximum allocation substantially. Pictorially, ---X--X--X--X--X-- those 5 X objects have limited future allocations to at most 6 dashes in size. Duncan Murdoch __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
On Tue, Oct 6, 2009 at 10:30 AM, Duncan Murdoch wrote: > When running a 32 bit program, 64 bit Windows hides most of itself outside > the address space visible to the program, so almost all of the 4 Gb address > space is available to the user. But no more: no matter how much RAM you > install, it's not possible to address it using a 32 bit address. That's true when running a 32-bit version of R under 64-bit Windows. But on the 64-bit version of REvolution R Enterprise for Windows, you can create objects much larger than 4Gb, for example: > memory.limit(2e+9) > v<-double(1e+9) # 8GB vector > sum(v) [1] 0 The same is true of 64-bit versions of R on other platforms, too. # David Smith -- David M Smith Director of Community, REvolution Computing www.revolution-computing.com Tel: +1 (206) 577-4778 x3203 (San Francisco, USA) Check out our upcoming events schedule at www.revolution-computing.com/events __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
On 10/6/2009 1:17 PM, Kjetil Halvorsen wrote: On Tue, Oct 6, 2009 at 11:23 AM, Duncan Murdoch wrote: On 10/6/2009 10:34 AM, Jose Quesada wrote: Robert Wilkins gmail.com> writes: Will R have more glitches on one operating system as opposed to another, or is it pretty much the same? robert One important difference is that, if you are unsing large datasets and need memory, then windows is by far the worst. CRAN R is 32 bit and can only address 1.5 Gb of memory (or something similar; I don't really understand why). By default, 32 bit Windows only gives 2 Gb for all the user processes to share, and saves the rest of memory for itself. You can change this (see the Windows FAQ), but the most you'll ever get is 3 Gb in 32 bit Windows, _and a bit under 4 Gb in 64 bit Windows. That sounds incredible. ¿Why so? When running a 32 bit program, 64 bit Windows hides most of itself outside the address space visible to the program, so almost all of the 4 Gb address space is available to the user. But no more: no matter how much RAM you install, it's not possible to address it using a 32 bit address. Duncan Murdoch Kjetil Duncan Murdoch While there's a 64-bit version of R for windows (revolution-computing.com) I would advise against using it, for several reasons. While revolution has provided very nice packages to the community (e.g., foreach), the win-64 port as of today is certainly the worst platform to do work on. Reasons: (1) it's R 2.7.2 (2) Many important packages will never be ported (3) Some packages (particularly those depending on Rjava) would not work properly (4) There's a proprietary repository, where most packages are outrageously outdated. (5) Most help you find on R-help will not apply. Instead, you have 'paid' support. Said support is slow, and close to useless in most cases. (6) Packages that rely on external tools (e.g., mysql) will take a lot of work to get going. And of course, one have to pay for a yearly license, to have the privilege to work under the above conditions. If you need 64-bit right now, my advice is to switch to basically any other platform. Note: this may change any time, since they are working on a continuous build that will keep the releases in sync with mainstream R. Jose Quesada, PhD. Max Planck Institute, Human Development, Berlin http://www.josequesada.name/ __ R-help@r-project.org mailing list 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. __ R-help@r-project.org mailing list 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. __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
On Tue, Oct 6, 2009 at 11:23 AM, Duncan Murdoch wrote: > On 10/6/2009 10:34 AM, Jose Quesada wrote: >> >> Robert Wilkins gmail.com> writes: >> >>> >>> Will R have more glitches on one operating system as opposed to >>> another, or is it pretty much the same? >>> >>> robert >>> >>> >> >> One important difference is that, if you are unsing large datasets and >> need >> memory, then windows is by far the worst. CRAN R is 32 bit and can only >> address 1.5 Gb of memory (or something similar; I >> don't really understand why). > > By default, 32 bit Windows only gives 2 Gb for all the user processes to > share, and saves the rest of memory for itself. You can change this (see > the Windows FAQ), but the most you'll ever get is 3 Gb in 32 bit Windows, > _and a bit under 4 Gb in 64 bit Windows. > That sounds incredible. ¿Why so? Kjetil > Duncan Murdoch > >> >> While there's a 64-bit version of R for windows (revolution-computing.com) >> I >> would advise against using it, for several reasons. While revolution has >> provided very nice packages to the community (e.g., foreach), the win-64 >> port as >> of today is certainly the worst platform to do work on. Reasons: >> (1) it's R 2.7.2 >> (2) Many important packages will never be ported >> (3) Some packages (particularly those depending on Rjava) would not work >> properly >> (4) There's a proprietary repository, where most packages are outrageously >> outdated. (5) Most help you find on R-help will not apply. Instead, you >> have 'paid' >> support. Said support is slow, and close to useless in most cases. >> (6) Packages that rely on external tools (e.g., mysql) will take a lot of >> work >> to get going. >> And of course, one have to pay for a yearly license, to have the privilege >> to >> work under the above conditions. >> >> If you need 64-bit right now, my advice is to switch to basically any >> other >> platform. >> >> Note: this may change any time, since they are working on a continuous >> build >> that will keep the releases in sync with mainstream R. >> >> Jose Quesada, PhD. >> Max Planck Institute, Human Development, Berlin >> http://www.josequesada.name/ >> >> __ >> R-help@r-project.org mailing list >> 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. > > __ > R-help@r-project.org mailing list > 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. > __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
On 10/6/2009 10:34 AM, Jose Quesada wrote: Robert Wilkins gmail.com> writes: Will R have more glitches on one operating system as opposed to another, or is it pretty much the same? robert One important difference is that, if you are unsing large datasets and need memory, then windows is by far the worst. CRAN R is 32 bit and can only address 1.5 Gb of memory (or something similar; I don't really understand why). By default, 32 bit Windows only gives 2 Gb for all the user processes to share, and saves the rest of memory for itself. You can change this (see the Windows FAQ), but the most you'll ever get is 3 Gb in 32 bit Windows, and a bit under 4 Gb in 64 bit Windows. Duncan Murdoch While there's a 64-bit version of R for windows (revolution-computing.com) I would advise against using it, for several reasons. While revolution has provided very nice packages to the community (e.g., foreach), the win-64 port as of today is certainly the worst platform to do work on. Reasons: (1) it's R 2.7.2 (2) Many important packages will never be ported (3) Some packages (particularly those depending on Rjava) would not work properly (4) There's a proprietary repository, where most packages are outrageously outdated. (5) Most help you find on R-help will not apply. Instead, you have 'paid' support. Said support is slow, and close to useless in most cases. (6) Packages that rely on external tools (e.g., mysql) will take a lot of work to get going. And of course, one have to pay for a yearly license, to have the privilege to work under the above conditions. If you need 64-bit right now, my advice is to switch to basically any other platform. Note: this may change any time, since they are working on a continuous build that will keep the releases in sync with mainstream R. Jose Quesada, PhD. Max Planck Institute, Human Development, Berlin http://www.josequesada.name/ __ R-help@r-project.org mailing list 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. __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
Robert Wilkins gmail.com> writes: > > Will R have more glitches on one operating system as opposed to > another, or is it pretty much the same? > > robert > > One important difference is that, if you are unsing large datasets and need memory, then windows is by far the worst. CRAN R is 32 bit and can only address 1.5 Gb of memory (or something similar; I don't really understand why). While there's a 64-bit version of R for windows (revolution-computing.com) I would advise against using it, for several reasons. While revolution has provided very nice packages to the community (e.g., foreach), the win-64 port as of today is certainly the worst platform to do work on. Reasons: (1) it's R 2.7.2 (2) Many important packages will never be ported (3) Some packages (particularly those depending on Rjava) would not work properly (4) There's a proprietary repository, where most packages are outrageously outdated. (5) Most help you find on R-help will not apply. Instead, you have 'paid' support. Said support is slow, and close to useless in most cases. (6) Packages that rely on external tools (e.g., mysql) will take a lot of work to get going. And of course, one have to pay for a yearly license, to have the privilege to work under the above conditions. If you need 64-bit right now, my advice is to switch to basically any other platform. Note: this may change any time, since they are working on a continuous build that will keep the releases in sync with mainstream R. Jose Quesada, PhD. Max Planck Institute, Human Development, Berlin http://www.josequesada.name/ __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
On 06/10/2009 6:36 AM, John Maindonald wrote: I had a large job time ago that ran fine under MacOS X. I'd expect the same to be true under Linux. It would run under Windows XP only if XP had been freshly rebooted. That sounds like you were running out of memory, or it was a bug. Did you report it? Duncan Murdoch John Maindonald email: john.maindon...@anu.edu.au phone : +61 2 (6125)3473fax : +61 2(6125)5549 Centre for Mathematics & Its Applications, Room 1194, John Dedman Mathematical Sciences Building (Building 27) Australian National University, Canberra ACT 0200. On 06/10/2009, at 9:00 PM, r-help-requ...@r-project.org wrote: From: Liviu Andronic Date: 6 October 2009 6:46:33 PM AEDT To: Robert Wilkins Cc: r-help@r-project.org Subject: Re: [R] R on Linux, and R on Windows , any difference in maturity+stability? On 10/6/09, Robert Wilkins wrote: Will R have more glitches on one operating system as opposed to another, Probably not. or is it pretty much the same? Depending on the complexity of the code, it is pretty much the same. I recently had a (relatively simple) group project, with the three of us on different OSs: Win, Mac and Linux. We did not encounter one platform specific issue. Liviu [[alternative HTML version deleted]] __ R-help@r-project.org mailing list 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. __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
I had a large job time ago that ran fine under MacOS X. I'd expect the same to be true under Linux. It would run under Windows XP only if XP had been freshly rebooted. John Maindonald email: john.maindon...@anu.edu.au phone : +61 2 (6125)3473fax : +61 2(6125)5549 Centre for Mathematics & Its Applications, Room 1194, John Dedman Mathematical Sciences Building (Building 27) Australian National University, Canberra ACT 0200. On 06/10/2009, at 9:00 PM, r-help-requ...@r-project.org wrote: > From: Liviu Andronic > Date: 6 October 2009 6:46:33 PM AEDT > To: Robert Wilkins > Cc: r-help@r-project.org > Subject: Re: [R] R on Linux, and R on Windows , any difference in > maturity+stability? > > > On 10/6/09, Robert Wilkins wrote: >> Will R have more glitches on one operating system as opposed to >> another, >> > Probably not. > >> or is it pretty much the same? >> > Depending on the complexity of the code, it is pretty much the same. I > recently had a (relatively simple) group project, with the three of us > on different OSs: Win, Mac and Linux. We did not encounter one > platform specific issue. > Liviu [[alternative HTML version deleted]] __ R-help@r-project.org mailing list 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.
Re: [R] R on Linux, and R on Windows , any difference in maturity+stability?
On 10/6/09, Robert Wilkins wrote: > Will R have more glitches on one operating system as opposed to > another, > Probably not. > or is it pretty much the same? > Depending on the complexity of the code, it is pretty much the same. I recently had a (relatively simple) group project, with the three of us on different OSs: Win, Mac and Linux. We did not encounter one platform specific issue. Liviu __ R-help@r-project.org mailing list 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.