Use FFLAGS for PARPACK with autotools and mpi #449
Conversation
| @@ -1,8 +1,6 @@ | |||
| AUTOMAKE_OPTIONS = subdir-objects # Needed as debug/stat* are not in current directory. | |||
|
|
|||
| F77 = $(MPIF77) | |||
| FFLAGS_SAV = @FFLAGS@ | |||
There was a problem hiding this comment.
@sylvestre: OK for you? Was this intentional?
There was a problem hiding this comment.
Yes this was intentional. I missed it is still used by libparpack@LIBSUFFIX@@ITF64SUFFIX@_la_FFLAGS = $(FFLAGS_SAV) which can now use FFLAGS instead. Or keep it to a one line change?
There was a problem hiding this comment.
libparpack@LIBSUFFIX@@ITF64SUFFIX@_la_FFLAGS = $(FFLAGS_SAV) which can now use FFLAGS instead.
I guess this is a good move.
Or keep it to a one line change?
@sylvestre: what do you think? Was this FFLAGS_SAV intentional (bug fix?) back in the old days?!
There was a problem hiding this comment.
I don't know, is there anything interesting in the git blame/log ? :)
There was a problem hiding this comment.
@turboencabulator: git-blame is on you! :)
Why did you need FFLAGS_SAV? Bug fix? If yes, which one? Can you review this PR?
There was a problem hiding this comment.
What problem was the no-optimization stuff trying to fix?
Don't know!
@loqs and @turboencabulator: I let you converge to the best solution to PR. Let me know when the PR is ready to be merged.
There was a problem hiding this comment.
@loqs: is there the same problem with cmake? I guess no.
Otherwise feel free to PR the equivalent fix at cmake side.
There was a problem hiding this comment.
2. What `FFLAGS` are you trying to add? Is it something that `configure` should be detecting and adding to `AM_FFLAGS`?
FFLAGS=-march=x86-64 -mtune=generic -O2 -pipe -fno-plt -fexceptions -Wp,-D_FORTIFY_SOURCE=2 -fstack-clash-protection -fcf-protection -ffile-prefix-map=/build/arpack/src=/usr/src/debug/arpack
I detected -fcf-protection was missing from PARPACK.
There was a problem hiding this comment.
I was thinking your FFLAGS might be related to MPI, but doesn't look like it.
Usually when I see -O0 it's because the code has some obscure bug that the optimizer exposes. But these two files seem fairly simple, and if CMake is building them optimized and nobody has complained yet... I say we remove the no-optimization stuff.
|
Closing in favor of #450. |
Fixes #448.