RAxML 64 bit version ???

[Adri-UCL]

Hi there,
I'm Adrien, postdoc at UCL.
I'm trying to use RAxML because I'm first interested in launching PartitionFinder through it.
Rob Lanfear (PartitionFinder) advised me to do the following:

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

1. Download and compile the latest version of raxml from here: https://github.com/stamatak/standard-RAxML (one of the Pthreads versions is best)

2. Rename the executable 'raxml' or 'raxml.exe' if you're on windows

3. Get the latest version of PF-raxml

4. Put your new 'raxml' program in the 'programs' folder of PartitionFinder

5. Run PartitionFinder with this extra option at the commandline: "--raxml"

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

I'm getting an error at the last step. I attached the error message as .jpg file. It looks like the RAxML version in not working with windows 7 64 bits.
By hazard, is there any version of RAxML that could allow me to run the PF analysis on windows 64b ?

Thanks
Adrien

[Alexandros Stamatakis]

Hi Adrien,

I am ccing Rob Lanfear on this.

Where did you download the windows version from?

We are not maintaining an up to date Win version,
to find the link for getting the most up to date RAxML windows version
please search for previous posts in the google group.

Apart from this your best choice would be to use Linux :-)

Alexis

n Wed, 2012-11-14 at 07:32 -0800, Adri-UCL wrote:
> Hi there,
> I'm Adrien, postdoc at UCL.
> I'm trying to use RAxML because I'm first interested in launching
> PartitionFinder through it.
> Rob Lanfear (PartitionFinder) advised me to do the following:
>
> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> 1. Download and compile the latest version of raxml from here:
> https://github.com/stamatak/standard-RAxML (one of the Pthreads versions is
> best)
>
> 2. Rename the executable 'raxml' or 'raxml.exe' if you're on windows
>

> 3. Get the latest version of PF-raxml<https://github.com/brettc/partitionfinder/tree/parallel_feature>

>
> 4. Put your new 'raxml' program in the 'programs' folder of PartitionFinder
>
> 5. Run PartitionFinder with this extra option at the commandline: "--raxml"
> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> I'm getting an error at the last step. I attached the error message as .jpg
> file. It looks like the RAxML version in not working with windows 7 64 bits.
> By hazard, is there any version of RAxML that could allow me to run the PF
> analysis on windows 64b ?
>
> Thanks
> Adrien
>

> <https://lh4.googleusercontent.com/-LQF7dYaOKoU/UKO5X02dkrI/AAAAAAAAAbk/og_OlMG1Pek/s1600/error.png>

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
Adjunct Professor, Dept. of Ecology and Evolutionary Biology, University
of Arizona at Tucson

www.exelixis-lab.org

[Adri-UCL]

Thanks for your answer Alexis.
Yes, in the meantime i managed to make it work using linux...
Adrien

[Rob Lanfear]

Hi All,

We have been working on this too. In the end it seems like it works on windows with gcc in cygwin, although the Pthreads version doesn't work. I suspect we need to get some extra libraries for that. 

Note that these instructions for using PartitionFinder are for an unreleased version. The code is public, so you (and others) should feel free to try it out. However, we will be releasing a more polished version in the near future, which will be available from the PartitionFinder website. 

Cheers,

Rob

--
Rob Lanfear
Research Fellow,
Ecology, Evolution, and Genetics,
Research School of Biology,
Australian National University

phone: +61 (0)2 6125 3611

www.robertlanfear.com

[Ingo Michalak]

Hey Rob and others,

the PTHREADS version work as well, when you just copy the missing .dll files from the cygwin folder to the folder where you run the RAxML exe.

At least this worked for me, when you use the mingw cross compiler version of gcc, that comes with cygwin. And I think at some point I had to hack the axml.c (it was just disabling an "if", which was related to POSIX...)... I can look it up later, if it is of interest.

Cheers

Ingo

[Akifumi S. Tanabe]

Hi Ingo and others,

I used MinGW-w64 compiler included in Strawberry Perl. All of
non-PTHREADS/i386, PTHREADS/i386, non-PTHREADS/x64, and PTHREADS/x64
versions work as well if "pthreadGC2-w??.dll" is placed in the same
folder, and if "-D__APPLE__" is added to CFLAGS. Because "posix_memalign"
is not supported by MinGW, I hope this function replaced to
"_aligned_malloc" on the Windows systems when possibe. Note that memory
reserved by "_aligned_malloc" must be released by "_aligned_free".

Best regards,

On Tue, 20 Nov 2012 05:23:21 -0800 (PST)
Ingo Michalak <boro...@gmail.com> wrote in
<107506e5-1940-4a58...@googlegroups.com>

> the PTHREADS version work as well, when you just copy the missing .dll
> files from the cygwin folder to the folder where you run the RAxML exe.
> At least this worked for me, when you use the mingw cross compiler version
> of gcc, that comes with cygwin. And I think at some point I had to hack the
> axml.c (it was just disabling an "if", which was related to POSIX...)... I
> can look it up later, if it is of interest.

--
Akifumi S. Tanabe, Ph.D.
Graduate School of Global Environmental Studies, Kyoto University,
Yoshida-Nihonmatsucho, Sakyo-ku, Kyoto, Kyoto, 606-8501, Japan.

[Ingo Michalak]

Dear Akifumi,

thanks for this. working with this flag is indeed much cleaner, than messing around with the c code. While trying if it also works with cygwin, I realized, that Alexis changed the code at exactly the bits where the workaround becomes important; so it doesn't work at all to compile PTHREADS versions from the new RAxML code (so Rob was right). But taking one of the axml.c files from before, the compilation works fine also with cygwin (using the x86_64-w64-mingw32-gcc.exe and changing "-pthread" to "-lpthread" in LIBRARIES, and as mentioned before, you have to copy the pthreadGC2.dll from the compiler to the path where the RAxML exe is located). 

@Alexis: is it inevitable to keep those changes? I know you are explicitely not providing Windows support, though...

Best

Ingo

Am Dienstag, 20. November 2012 17:00:56 UTC+1 schrieb Akifumi:

Hi Ingo and others,

  I used MinGW-w64 compiler included in Strawberry Perl. All of
non-PTHREADS/i386, PTHREADS/i386, non-PTHREADS/x64, and PTHREADS/x64
versions work as well if "pthreadGC2-w??.dll" is placed in the same
folder, and if "-D__APPLE__" is added to CFLAGS. Because "posix_memalign"
is not supported by MinGW, I hope this function replaced to
"_aligned_malloc" on the Windows systems when possibe. Note that memory
reserved by "_aligned_malloc" must be released by "_aligned_free".

Best regards,

On Tue, 20 Nov 2012 05:23:21 -0800 (PST)
Ingo Michalak <boro...@gmail.com> wrote in

<107506e5-1940-4a58-a723-e2c691...@googlegroups.com>

[Simon Berger]

Hi All,

  I used MinGW-w64 compiler included in Strawberry Perl. All of
non-PTHREADS/i386, PTHREADS/i386, non-PTHREADS/x64, and PTHREADS/x64
versions work as well if "pthreadGC2-w??.dll" is placed in the same
folder, and if "-D__APPLE__" is added to CFLAGS. Because "posix_memalign"

I am maintaining a branch on github that should compile out of the box on mingw (see https://github.com/sim82/standard-RAxML/branches it's the win32 branch). I have just updated it to the latest raxml version in Alexi's github repository.

WARNING: The new FASTA reader in raxml requires a function (getline) that is not available in mingw, so I have just implemented it. This code is _completely_ untested (i.e., I do not even know if the FASTA reader works at all. I've tried to make it as correct as possible, but text-file input is one of the nastiest things possible, not only if you are flying blind.  I'll try to test it tomorrow).

 

is not supported by MinGW, I hope this function replaced to
"_aligned_malloc" on the Windows systems when possibe. Note that memory
reserved by "_aligned_malloc" must be released by "_aligned_free".

this is only a problem for the SSE3 version. The _aligned_free has been a very effective show-stopper so far, and I could not figure out a good solution that is not a complete hack...

Regards,

Simon

 

Best regards,

On Tue, 20 Nov 2012 05:23:21 -0800 (PST)
Ingo Michalak <boro...@gmail.com> wrote in

<107506e5-1940-4a58-a723-e2c691...@googlegroups.com>

[Ingo Michalak]

Thank you Simon!

I just tried: it also works from cygwin, when using the mingw compilers and changing to lpthread library (see above).

Cheers

Ingo

[Alexandros Stamatakis]

Many thanks to everybdoy,

the change I applied is inevitable, it was a fix for compiling the AVX
code on MACs with clang :-(

Simon, could you send me your getline() implementation please?

I'll replace the GNU version with yours and test it.

Cheers,

Alexis

> >> <107506e5-1940-4a58...@googlegroups.com>

> >> > the PTHREADS version work as well, when you just copy the missing .dll
> >> > files from the cygwin folder to the folder where you run the RAxML exe.
> >> > At least this worked for me, when you use the mingw cross compiler
> >> version
> >> > of gcc, that comes with cygwin. And I think at some point I had to hack
> >> the
> >> > axml.c (it was just disabling an "if", which was related to
> >> POSIX...)... I
> >> > can look it up later, if it is of interest.
> >>
> >>
> >> --
> >> Akifumi S. Tanabe, Ph.D.
> >> Graduate School of Global Environmental Studies, Kyoto University,
> >> Yoshida-Nihonmatsucho, Sakyo-ku, Kyoto, Kyoto, 606-8501, Japan.
> >>
> >>

[Ingo Michalak]

Dear Alexis,

the compilation of 32 and 64 bit versions (PTHREADS and normal) for Windows works using Simon's adopted code, so it's all good :)

(and as far as I understood Simon it was not the code piece I mentioned, that caused the issue, but rather the FASTA parsing)

Best

Ingo

> >> <107506e5-1940-4a58-a723-e2c691...@googlegroups.com>

[Simon Berger]

Hi Alexi,

Simon, could you send me your getline() implementation please?

I'll replace the GNU version with yours and test it.

it is in https://github.com/sim82/standard-RAxML/blob/win32/axml.c lines 608-667 (I'm posting this here in the hope that anyone who feels like it has a look to spot all the inevitable errors...). The buffer resize strategy in getline_insptr_valid could probably be optimized for files with very long lines (maybe some super-linear growth, or growth by full pages etc. Lots of research has been done on this.).

Simon

 

> >> <107506e5-1940-4a58-a723-e2c691...@googlegroups.com>

[Simon Berger]

it is in https://github.com/sim82/standard-RAxML/blob/win32/axml.c lines 608-667 (I'm posting this here in the hope that anyone who feels like it has a look to spot all the inevitable errors...). The buffer resize strategy in getline_insptr_valid could probably be optimized for files with very long lines (maybe some super-linear growth, or growth by full pages etc. Lots of research has been done on this.).

I've just pushed an update fixing two bugs. One was a slight misunderstanding of the specification, the second was too embarrassing to talk about (remember: do NEVER release code without any testing if you don't want to look like a complete fool...)

Simon 

[Nataly Nunes Slivak]

Hi, 

I can't execute RAxML for Windows 7 32bits. I tested "WindowsExecutables_v8.2.4" and "WindowsExecutables_v8.2.10" and not worked.

Nataly

[Alexandros Stamatakis]

Hi Nataly,

As we are not officially maintaining the windows version of the code I
can't help you with this.

I have bcced the two kind people who have been providing the WIN
executables for RAxML in the past, maybe they can help.

Alexis

On 29.09.2017 22:27, Nataly Nunes Slivak wrote:
> Hi,
>
> I can't execute RAxML for Windows 7 32bits. I tested
> "WindowsExecutables_v8.2.4" and "WindowsExecutables_v8.2.10" and not worked.
>
> Nataly
>
> Em quarta-feira, 14 de novembro de 2012 13:32:12 UTC-2, Adri-UCL escreveu:
>
> Hi there,
> I'm Adrien, postdoc at UCL.
> I'm trying to use RAxML because I'm first interested in launching
> PartitionFinder through it.
> Rob Lanfear (PartitionFinder) advised me to do the following:
>
> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> 1. Download and compile the latest version of raxml from here:
> https://github.com/stamatak/standard-RAxML
> <https://github.com/stamatak/standard-RAxML> (one of the Pthreads
> versions is best)
>
> 2. Rename the executable 'raxml' or 'raxml.exe' if you're on windows
>
> 3. Get the latest version of

> PF-raxml<https://github.com/brettc/partitionfinder/tree/parallel_feature>

>
> 4. Put your new 'raxml' program in the 'programs' folder of
> PartitionFinder
>
> 5. Run PartitionFinder with this extra option at the commandline:
> "--raxml"
>
> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> I'm getting an error at the last step. I attached the error message
> as .jpg file. It looks like the RAxML version in not working with
> windows 7 64 bits.
> By hazard, is there any version of RAxML that could allow me to run
> the PF analysis on windows 64b ?
>
> Thanks
> Adrien
>

> <https://lh4.googleusercontent.com/-LQF7dYaOKoU/UKO5X02dkrI/AAAAAAAAAbk/og_OlMG1Pek/s1600/error.png>
>
> --
> You received this message because you are subscribed to the Google
> Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to raxml+un...@googlegroups.com
> <mailto:raxml+un...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology

www.exelixis-lab.org