Hi, Ana.
You can expect gfortran-generated executables to run slower than the Intel-generated ones. I just ran a comparison on my PC: test 01 takes about 3.6 seconds (Intel) compared to 10.8 seconds (gfortran), even with similar optimization options.
However, looking at your output, I see some debugging information. It looks like you’ve compiled with options to give you run-time checks. This will certainly slow down the code. So, the first thing I would do is look at your FAST makefile and see what you’ve specified for FFLAGS. This is what I generally specify for production-type code:
FFLAGS = -O2 -m$(BITS) -fbacktrace -ffree-line-length-none -x f95-cpp-input
LDFLAGS = -O2 -m$(BITS) -fbacktrace
Thank you very much, Bonnie.
With you recommendation I get a two times quicker result even if it is not as fast as yours:
Hi, Ana.
I am having the exact same problem as you, where the project finds the lapack/blas library, but then gives the “Skipping incompatible…” message. I tried compiling the libraries myself, following the instructions at icl.cs.utk.edu/lapack-for-windows/lapack/, but I keep running into problems I do not understand at all.
Do you think you could help me out by uploading the 64 bit lapack/blas libraries you compiled with gfortran?
Thanks,
Mohibb Malik
Hi,
I have successfully compiled Fast v8 and Turbsim v2 using a gfortran compiler, and both are working very well. A strange issue has come to my attention, however.
I was getting errors such as ‘Segmentation Error- core dumped’ and found the reason for this error to be the “Hdsum” flag in the OC3Hywind HydroDyn.dat file. When this flag is set to ‘True’ - no errors. When it is set to ‘False’ - I get the segmentation error.
Could anyone explain why this might be happening?
Thanks,
Luke
Dear Luke,
Thanks for reporting this issue. Rafael Mudafort traced this back to a minor bug in HydroDyn, which he has now fixed in OpenFAST–see: github.com/OpenFAST/openfast/pull/207.
Best regards,
Hi Jason,
Great, thanks! I am now trying to compile FAST on a high performance computer using Linux to run some extensive simulations. I am running into some compilation error with FAST which I do not understand.
I’m trying to follow the instructions at github.com/old-NWTC/FAST/blob/m … owsCompile but running into a problem. I have the Registry and MAP++ steps done. I did have to edit the make file for how MAP++ links to lapacke but the library was generated so I don’t think that is the issue.
The problem I encounter is when doing a make in ~/src/FAST/Compiling leads to the following error:
Registry warning: type MeshMapType used before defined for IceD_P_2_SD_P
Registry warning: type MeshMapType used before defined for SD_P_2_IceD_P
generating …/Source/FAST_Types.f90
make: *** […/Source/FAST_Types.f90] Segmentation fault
make: *** Deleting file `…/Source/FAST_Types.f90’
I’ve attached the full output from the attempted compilation here.
Here are some environment details:
$ gfortran --version
GNU Fortran (GCC) 7.3.0
Copyright (C) 2017 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
$ gcc --version
gcc (GCC) 7.3.0
Copyright (C) 2017 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
$ cat /etc/redhat-release
Scientific Linux release 6.10 (Carbon)
$ uname -r
2.6.32-754.3.5.el6.x86_64
I’d really appreciate any help you can offer and sorry if I’m missing something basic or obvious.
Thanks,
Luke Cunningham
Dear Luke,
I’m not sure, but this old-NWTC version of FAST is no longer supported. Hopefully switching to OpenFAST will solve your problem: github.com/OpenFAST/openfast.
Best regards,
