Andrew M. Bradley
Andrew M. Bradley
> -fp-model=consistent is the most safe, but I think is only available for ifort (well the man page indicated ifort only, but icpx accepts it). I can try adjusting flags...
Oh, I wasn't thinking when I wrote "ifx" above. On pm-cpu, it's the C++ compiler that matters, as all the dycore code is in C++. Chrysalis uses icpc (ICC) 19.1.3.304...
Tagging others who might participate in this discussion: @PeterCaldwell @mt5555 @bartgol @AaronDonahue
Jim, would you point to an example of this pattern in the code? Thanks.
Jim, would you point me to your branch? I'm curious if all of the cases involve clipping with min and max.
One of my concerns here is it's not clear to me the original F90 is quite what we want. Then again, I might just be confused. Here is the original...
This case seems to have the same issue as above. If testvar = 0, then r_qwthl_1 = 0. But if it deviates from 0 by a tiny bit, then r_qwthl_1...
@bartgol I'm not convinced yet that code that properly implements the evaluation of a continuous function will necessarily require delayed evaluation to look cleaner. Re: the question you just asked,...
Luca, I don't disagree and have no opinion re: Jim's macro, your suggestions, etc. Obviously please feel free to introduce those additional mechanisms. My point is only that in the...
I'll take the easy one, 3. Just define it as hard-coded double rather than Real. Do that consistently in F90 and C++ when needing BFB tests. If implicit type conversion...