https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85531
Bug ID: 85531
Summary: Implement some loop fusion in the Fortran front end
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: enhancement
Priority: P3
Component: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: tkoenig at gcc dot gnu.org
Target Milestone: ---
See the discussion starting at https://gcc.gnu.org/ml/gcc/2018-04/msg00128.html
.
Basically, for code like
ZQICE(1:NPROMA,1:NFLEVG) = PGFL(1:NPROMA,1:NFLEVG,YI%MP)
ZQLI(1:NPROMA,1:NFLEVG) = PGFL(1:NPROMA,1:NFLEVG,YL%MP)
ZQRAIN(1:NPROMA,1:NFLEVG) = PGFL(1:NPROMA,1:NFLEVG,YR%MP)
ZQSNOW(1:NPROMA,1:NFLEVG) = PGFL(1:NPROMA,1:NFLEVG,YS%MP)
the front end could notice that the right-hand sides always agree
and translate this to
DO J=1,NFLEVG
DO I=1,NPROMA
ZQICE(I,J) = PGFL(I,J,YI%MP)
ZQLI(I,J) = PGFL(I,J,YI%MP)
ZQRAIN(I,J) = PGFL(I,J,YI%MP)
ZQSNOW(I,J) = PGFL(I,J,YI%MP)
END DO
END DO
This would require some data flow analysis (are we sure that PGFL
is not changed in the meantime?) and some basic block analysis.
