Compiling instruction #5
Comments
|
It does compile with GNU Fortran (gfortran). That being said, I needed to modify a few files. Mostly the OPEN commands were non standard. Additionaly you need to set switches to turn off bounds checking, accept Hollerith notation, and implicitly apply the SAVE command. Compiler switches I needed: You can check out my fork: The makefile is in the top level directory. It uses gfortan, and assumes that you have python in your $PATH. I did it in a linux environment, and it should work in a cygwin environment with minimal fuss. Cheers. |
|
I am sure this is not the place to be asking this – but I don’t know where else to go. I tried contacting NASA, but was not successful in getting any answers. So, forgive me – where can a non-programmer (like me) get an executable copy of NASTRAN-95?
Does anyone offer (or sell) such a thing? I would like to run NASTRAN-95 on a XP machine. Any suggestions would be greatly appreciated.
Thanks,
Dave
From: Dan [mailto:[email protected]]
Sent: Monday, March 13, 2017 2:04 PM
To: nasa/NASTRAN-95 <[email protected]>
Cc: Subscribed <[email protected]>
Subject: Re: [nasa/NASTRAN-95] Compiling instruction (#5)
It does compile with GNU Fortran (gfortran). That being said, I needed to modify a few files. Mostly the OPEN commands were non standard. Additionaly you need to set switches to turn off bounds checking, accept Hollerith notation, and implicitly apply the SAVE command.
Compiler switches I needed:
-fno-range-check -fno-automatic -std=legacy
You can check out my fork:
https://github.com/AeroDME/NASTRAN-95
The makefile is in the top level directory. It uses gfortan, and assumes that you have python in your $PATH. I did it in a linux environment, and it should work in a cygwin environment with minimal fuss.
Cheers.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub <#5 (comment)> , or mute the thread <https://github.com/notifications/unsubscribe-auth/AZEUwNs7Sqh4obMLqaBLeUCwwaVBWWFDks5rlYT-gaJpZM4MbmTN> .
|
|
Hi Dave, Yeah, this requires a little knowledge of compiling. You could read up on CYGWIN which is a linux environment that runs on windows. I am sure a windows fortran compiler will do it, and you can use the makefile specified prevously. But I don't know where you can get just the binaries. In fact, I am not sure the NASA license permits distribution of the binaries. Also, make sure this is what you want. You cannot necessarily take a MSC/NX NASTRAN bulk data file and just run it. It doesn't know what SOL 101 OR NLSTATIC means. It is not modern. Cheers. |
|
Dan – thanks for your reply – I guess I will stay away from this adventure!! I have used MSC/NASTRAN for years and thought perhaps I could get something close to MSC without the huge cost. I guess I was dreaming!!
Thanks again for your insight,
Dave
From: Dan [mailto:[email protected]]
Sent: Monday, March 13, 2017 9:36 PM
To: nasa/NASTRAN-95 <[email protected]>
Cc: dhhudson <[email protected]>; Comment <[email protected]>
Subject: Re: [nasa/NASTRAN-95] Compiling instruction (#5)
Hi Dave,
Yeah, this requires a little knowledge of compiling. You could read up on CYGWIN which is a linux environment that runs on windows. I am sure a windows fortran compiler will do it, and you can use the makefile specified prevously. But I don't know where you can get just the binaries. In fact, I am not sure the NASA license permits distribution of the binaries.
Also, make sure this is what you want. You cannot necessarily take a MSC/NX NASTRAN bulk data file and just run it. It doesn't know what SOL 101 OR NLSTATIC means. It is not modern.
Cheers.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub <#5 (comment)> , or mute the thread <https://github.com/notifications/unsubscribe-auth/AZEUwCGafAnh0KC2hlMPOrDqqd0NA0TCks5rle7ygaJpZM4MbmTN> .
|
|
FYI, this is packaged in Debian: And the sources contain a autotools build system: https://github.com/ldallolio/NASTRAN-95 by @ldallolio |
|
I was able to build twuerfl/NASTRAN-95 on mac os 10.12 but it segfaults when I run one of the test input files. So I think I will give yours a try. Do you think it will work? |
|
I think it is worth a shot. I have run it successfully on ARM and AMD
processors. The nastran.py script should give you some clues as to setting
environment variables and such.
It generates a bash script, so you may want to play with it a bit.
I was getting a segfault because it was not setting up memory correctly and
most of the issues seem to be resolved by setting up a bunch of environment
variables.
cheers!
Dan
…On Tue, Dec 18, 2018 at 5:48 PM mbrethen ***@***.***> wrote:
I was able to build twuerfl/NASTRAN-95 on mac os 10.12 but it segfaults
when I run one of the test input files.
==== NASTRAN is beginning execution of bar ====
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
#0 0x16746bf8f
#1 0x16746b772
#2 0x7fffb1879b39
#3 0x1079eb44f
#4 0x1079eb232
#5 0x1079e5fea
#6 0x1078d4273
#7 0x1079fb9c4
#8 0x1077f997f
#9 0x107a13b1d
Segmentation fault
So I think I will give yours a try. Do you think it will work?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#5 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AQUZQzMQSCSA0MmPhOSxF-hHwh0Xsm3sks5u6X6zgaJpZM4MbmTN>
.
|
|
I intend to install everything into /opt/local/... I'm aware that the nastran executable has hardcoded paths, e.g. Thanks |
|
It is expecting the file tree to be as it is downloaded. So, after you
compile it should be ready to copy to your final location. Unfortunately,
I have not made anything to install it.
So, in the sbin/nastran.py script, the following line sets the rf directory:
RFDIR = '{0}/rf' .format(nastran_home)
You could set that as you please to be a different value. Or comment it
out and set the RFDIR environment variable manually before your nastran run.
If you do comment it out, you will want to also comment out the following
line that actually creates the job script:
jobscr += '\nexport RFDIR={0}'.format(RFDIR)
Another thing to note. If you run the nastran.py script, it has a --help
option to give all the options. One option -n (--no-run) will take the all
the inputs but will generate the job script and print it to STDOUT instead
of running it. That way, you can see what is having to be set up to run
the job.
Does that make sense?
Cheers,
Dan
…On Thu, Dec 20, 2018 at 8:44 AM mbrethen ***@***.***> wrote:
I intend to install everything into /opt/local/... I'm aware that the
nastran executable has hardcoded paths, e.g. set
rfdir=/usr/users/bob/nast95/rf and I'll need to patch these. What is the
expected directory tree layout?
Thanks
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#5 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AQUZQ5dZEJRtdO1QxkvgnHYnApEbahLZks5u66I1gaJpZM4MbmTN>
.
|
|
I didn't recognize the string replacement at first. I had to look at my python reference to see that you're using string replacement with explicit numbering (I'm used to %(foo)s ), setting RFDIR = nastran_home/rf. nastran_home is set to the absolute path where sbin/nastran.py resides. |
|
I tried to build it this morning without success. Asside from several warnings of the type:
It returned with an error: This usually means that libnas.a is compiled for 32-bit, but the rest of the project is compiled for 64-bit. |
|
If you want to build the Debian/Ubuntu package, the method is super easy |
|
According to this discussion I patched the makefile and used libtool to get it to compile. So, after building I end up with: It appears the OUTPUT directory contains test runs. Looking at one of the f06 files: I guess this is because I'm building in /opt/local/... whereas the user accounts are in /Users/... on macs? |
|
For the moment, I have installed these directories in /opt/local/libexec/NASTRAN-95: I've set nastran_home to this directory and created a symlink to sbin/nastran.py in /opt/local/bin, which is in my PATH. I ran a simple bar input file and the f06 indicated it finished successfully. bin/nasthelp.x won't load the .TXT files (says they don't exist). The text files are in the bin directory. Have you used nastplot.x, chkfil.x or ff.x? What are the fortran files in the obsolete folder? |
|
Building nastran-95 is not for the faint of heart; and at the end of the day, assuming you are persistent, you will end up with a obsolete code. That being said I have had success using CODE BLOCKS with ubuntu linux 18.04. ( Ubuntu can also be run in Win 10.) I created several fortran library projects as well as the main application finally produced a running executable. The biggest plus is that it incorprates the DMAP compiler and runs the example problems that come with the release. The element library and the solvers are woefully obsolete: it represents 30 year old technology. However, if you have the time and expertise you can use what I call the NOS (Nastran Operating System) to create a modern up-to-date system. Think of combining JuliaFEM element modules in a rather elegant nastran system. I don't want to get bogged down with raising 95 from the library of dead codes but would be happy to respond to requests from those in academia and consultants like Dave for a copy of my code that I call CML (Computational Mechanics Laboratory). I believe you will be impressed with it. Contact me if you are interested. |
|
Hi Harry – I am definitely interested! As you know, I am not a “coder” or programmer, so anything that would run in Windows 10 would be of interest to me. Please send me the details, including cost.
Dave
From: Harry schaeffer [mailto:[email protected]]
Sent: Tuesday, August 20, 2019 5:58 PM
To: nasa/NASTRAN-95 <[email protected]>
Cc: dhhudson <[email protected]>; Comment <[email protected]>
Subject: Re: [nasa/NASTRAN-95] Compiling instruction (#5)
Building nastran-95 is not for the faint of heart; and at the end of the day, assuming you are persistent, you will end up with a obsolete code. That being said I have had success using CODE BLOCKS with ubuntu linux 18.04. ( Ubuntu can also be run in Win 10.) I created several fortran library projects as well as the main application finally produced a running executable.
The biggest plus is that it incorprates the DMAP compiler and runs the example problems that come with the release. The element library and the solvers are woefully obsolete: it represents 30 year old technology. However, if you have the time and expertise you can use what I call the NOS (Nastran Operating System) to create a modern up-to-date system. Think of combining JuliaFEM element modules in a rather elegant nastran system.
I don't want to get bogged down with raising 95 from the library of dead codes but would be happy to respond to requests from those in academia and consultants like Dave for a copy of my code that I call CML (Computational Mechanics Laboratory). I believe you will be impressed with it. Contact me if you are interested.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub <#5?email_source=notifications&email_token=AGIRJQDKMILDLFD57AEPQJLQFRSH7A5CNFSM4DDOMTG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4XZE2A#issuecomment-523211368> , or mute the thread <https://github.com/notifications/unsubscribe-auth/AGIRJQHABLY6XMC2EACU5GLQFRSH7ANCNFSM4DDOMTGQ> . <https://github.com/notifications/beacon/AGIRJQEKLDSNXD52LFXMIM3QFRSH7A5CNFSM4DDOMTG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4XZE2A.gif>
|
|
Hi Dave,
Have you tried to implement the WSL on win 10? Perhaps that is a bridge too
far. As an alternative I do have a windows version that runs on win 10.
This would be a better alternative at this time. Let me know.
Best,
Harry
…On Tue, Aug 20, 2019 at 3:37 PM dhhudson ***@***.***> wrote:
Hi Harry – I am definitely interested! As you know, I am not a “coder” or
programmer, so anything that would run in Windows 10 would be of interest
to me. Please send me the details, including cost.
Dave
From: Harry schaeffer ***@***.***
Sent: Tuesday, August 20, 2019 5:58 PM
To: nasa/NASTRAN-95 ***@***.***>
Cc: dhhudson ***@***.***>; Comment <
***@***.***>
Subject: Re: [nasa/NASTRAN-95] Compiling instruction (#5)
Building nastran-95 is not for the faint of heart; and at the end of the
day, assuming you are persistent, you will end up with a obsolete code.
That being said I have had success using CODE BLOCKS with ubuntu linux
18.04. ( Ubuntu can also be run in Win 10.) I created several fortran
library projects as well as the main application finally produced a running
executable.
The biggest plus is that it incorprates the DMAP compiler and runs the
example problems that come with the release. The element library and the
solvers are woefully obsolete: it represents 30 year old technology.
However, if you have the time and expertise you can use what I call the NOS
(Nastran Operating System) to create a modern up-to-date system. Think of
combining JuliaFEM element modules in a rather elegant nastran system.
I don't want to get bogged down with raising 95 from the library of dead
codes but would be happy to respond to requests from those in academia and
consultants like Dave for a copy of my code that I call CML (Computational
Mechanics Laboratory). I believe you will be impressed with it. Contact me
if you are interested.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub <
#5?email_source=notifications&email_token=AGIRJQDKMILDLFD57AEPQJLQFRSH7A5CNFSM4DDOMTG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4XZE2A#issuecomment-523211368>
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/AGIRJQHABLY6XMC2EACU5GLQFRSH7ANCNFSM4DDOMTGQ>
. <
https://github.com/notifications/beacon/AGIRJQEKLDSNXD52LFXMIM3QFRSH7A5CNFSM4DDOMTG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4XZE2A.gif>
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#5?email_source=notifications&email_token=ACEVYDOI2MVOHDFZQWJXXYDQFRWZVA5CNFSM4DDOMTG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4X3XII#issuecomment-523221921>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ACEVYDMHK2PV53ZVXEOK6X3QFRWZVANCNFSM4DDOMTGQ>
.
--
Harry Schaeffer
9 Camino Botero
San Clemente, CA 92673
cell: 949-547-4661
tele: 949-940-0211
|
|
Hello all. Can someone help me to compile that or have a compiled example to follow? I'm on Ubuntu.
|
|
I can help. I will respond in detail on Monday.
Harry
…On Sat, Apr 4, 2020, 8:24 AM diegomagela ***@***.***> wrote:
Hello all.
Can someone help me to compile that or have a compiled example to follow?
I'm on Ubuntu.
For example:
1. Do that
2. Than do that
3. ...
.
.
.
N. Finsh
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#5 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ACEVYDOS7AQRKKA6C4EM7WDRK5GJ7ANCNFSM4DDOMTGQ>
.
|
|
Hello,
Interesting that I have been working on an article that describes how to
build Calculix using Code Blocks. I work was done using Ubuntu 18.04 that
you mentioned in your note.
I have also used C::B to build a much enhanced version of Nastran95 that I
call ComLab. The architecture of Nastran is a keeper, with some major
improvements, but the FEA libraries are not useful in today's world. The 95
code has 2 main libraries: MIS are Machine independent source; and MDS are
Machine dependent source. You might try these first using C::B.
It you are looking for a modern Nastran, there is always MSC.Nastran from
MSC Software. I also offer ComLab which is delivered on a memory stick for
$100 that covers packaging and shipping. If this is of interest I will send
you a copy of the EULA.
I would note that I can't fulfill any orders until the current crisis is
over. Oh, I am attaching a PDF of the article, "Building Calculix with Code
Blocks ". ( I apologize for typos!)
Best,
Harry
On Sat, Apr 4, 2020 at 11:51 AM Harry Schaeffer <[email protected]>
wrote:
… I can help. I will respond in detail on Monday.
Harry
On Sat, Apr 4, 2020, 8:24 AM diegomagela ***@***.***> wrote:
> Hello all.
>
> Can someone help me to compile that or have a compiled example to follow?
> I'm on Ubuntu.
> For example:
>
> 1. Do that
> 2. Than do that
> 3. ...
> .
> .
> .
> N. Finsh
>
> —
> You are receiving this because you commented.
> Reply to this email directly, view it on GitHub
> <#5 (comment)>, or
> unsubscribe
> <https://github.com/notifications/unsubscribe-auth/ACEVYDOS7AQRKKA6C4EM7WDRK5GJ7ANCNFSM4DDOMTGQ>
> .
>
--
Harry Schaeffer
9 Camino Botero
San Clemente, CA 92673
cell: 949-547-4661
tele: 949-940-0211
|
|
On Mon, Apr 6, 2020 at 12:32 PM diegomagela ***@***.***> wrote:
Harry, I thought you forgot the PDF.
—
Send me your email address
Harry
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#5 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ACEVYDOEVOR6HCDWQBEM4ELRLIU4RANCNFSM4DDOMTGQ>
.
--
Harry Schaeffer
9 Camino Botero
San Clemente, CA 92673
cell: 949-547-4661
tele: 949-940-0211
|
|
Hello everybody, We have some of the MYSTRAN, pyNASTRAN , and FreeCAD folks at the NASTRAN Discord channel working to get this going. It would be great to have you over there as well. let's see if we can revive this initiative. |
|
For those who end up here looking for ComLab, @harryschaeffer has now released ComLab as a freemium software. More information here |
|
I make a dir called opt under /home, with copying nastran-95-master,and building it ,and I success,but when I test the job d010000a.inp I get a fatal " RFOPEN CAN NOT OPEN /public1/home/sc91522/opt/NASTRAN-95-master/", can anyone help me recorrect this issue? added that my system is contos 7.9 x64 |
|
Do u mean that I must build it under the root dir ? |
It seems that 'SMCOMX.COM' and 'MMACOM.COM' are necessary files in the 'mis' folder, but these two files are missing in the current version. |
|
OK，thanks for ur work发自我的 iPhone在 2024年7月9日，13:43，woodifsh ***@***.***> 写道：
It does compile with GNU Fortran (gfortran). That being said, I needed to modify a few files. Mostly the OPEN commands were non standard. Additionaly you need to set switches to turn off bounds checking, accept Hollerith notation, and implicitly apply the SAVE command.
Compiler switches I needed:
-fno-range-check -fno-automatic -std=legacy
You can check out my fork:
https://github.com/AeroDME/NASTRAN-95
The makefile is in the top level directory. It uses gfortan, and assumes that you have python in your $PATH. I did it in a linux environment, and it should work in a cygwin environment with minimal fuss.
Cheers.
It seems that 'SMCOMX.COM' and 'MMACOM.COM' are necessary files in the 'mis' folder, but these two files are missing in the current version.
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: ***@***.***>
|
Can someone explain how to compile this marvel? Will it work on a PC compiled with GNU Fortran? There are no makefile nor instructions for compiling. Thanks in advance!
The text was updated successfully, but these errors were encountered: