Compiling old version of FAST (6.10) with IVF

Wind & Water Computer-Aided Engineering Software Tools
1

Hello,

I’m in a bit of a pickle, so I’m asking for advice. I’m trying to compile FAST v6.10 in Intel Visual Fortran (IVF), for reasons explained below, and I’ve had various compiling, run time, and output errors. Currently my output for a wind ramp is NaN on most parameters. Any clues would help. I see from the change log that you did not start compiling with IVF until you had developed the NWTC Subs.

The reason I’m doing this is that I developed code for swept blades (CurveFAST) using FAST v6.10. Lately I’ve been compiling in gfortran but users have been complaining of excessive runtimes- taking as long as a similar Adams run. I finally purchased IVF, and I want to start with a baseline build of FAST v6.10 to compare runtimes with a similar model in CurveFAST compiled in IVF.

Thank you for your time,

2

Scott,

At one point, we had a list of necessary changes to compile FAST in IVF:

I do not think step (2) is necessary, and I seem to recall reading that later versions of IVF use shortcutting (?) so (5) may not be necessary.

Good luck!

3

Great-

I’ve seen almost all these errors come up one-by-one. I will post if I find anything else.

4

Hi Bonnie,

I’m making progress…

First I wanted to say that I’m actually working from FAST v6.00, 10-Jun-2005, no 6.10 as stated previously.

Sysifort did not work, I had to modify SysCVF to look more like SysIVF in the NWTC library. This file is attached for the record.

I have the executable running a wind ramp for the WindPACT 1.5. Everything looks fine EXCEPT the TipRxc1,2,3 are printed out as “**********” with an occasional “0.000E+00” at near the end of the ramp. When running with old FAST executable, I get reasonable numbers for this deflection. Please let me know if you have any idea what might be happening to these deflections.

Thanks,
Scott Larwood
SysIVF.txt (7.6 KB)

5

I found out that blade deflections were bogus with /O3 optimization (full and higher level), but work with /O2 optimization (maximize speed). I found this after turning off the optimization for debugging. I’m moving on from here, just using the /O2 optimization.