If I compile the following program with
ifort -m32 -heap-arrays test.f
the array in mytest2() is allocated on the heap, but the one in mytest() is not. According to the documentation, the -heap-arrays flag should put all automatic arrays on the heap. Am I misunderstanding the documentation? Is there a way to force the array in mytest to be allocated on the heap instead of on the stack or in static data?
write(CSTmp,"(I0)") size(this%RVPEffSize) CSFormat="("//Trim(AdjustL(CSTmp))//"(1X,F20.10))" write(this%ISOutputUnitEff,CSFormat,iostat=ISSubStat,iomsg=CSMsg)& &(this%RVPEffSize(c1),c1=1,size(this%RVPEffSize))
is supposed to write the content of "this%RVPEffSize" into a single line of the file. It does so until "size(this%RVPEffSize)" is 51422. For larger sizes, a new line is inserted.
Is that intended or a bug or related to the operation system???
ifort 14, ubuntu 14.04
I downloaded the trial version of Composer XE 2015.1.133. The webpage said it also included fortran compiler. I am unable to locate ifp or ifort anywhere in the package. Am I overlooking something obvious? Or is the compiler called something else? Thanks for any hints.
I am just playing with integer overflow in fortran compilers and have found that this simple "buggy" fortran code causes internal compiler error by ifort on each of the write instruction. Gfortran detects these overflows and shows some "Error: Arithmetic overflow" during compilation.
I compiled this file with following commands:
And the error is:
fortran_2003_ieee_test.F(39): error #6503: This keyword is invalid for this intrinsic procedure reference. [UNIT]
CALL FLUSH (UNIT=6)
compilation aborted for fortran_2003_ieee_test.F (code 1)
Here's the code:
USE , INTRINSIC :: IEEE_EXCEPTIONS
$ ifort --version ifort (IFORT) 15.0.0 20140723 Copyright (C) 1985-2014 Intel Corporation. All rights reserved.
The following Fortran code:
I am having some trouble trying to adapt some fortran code to BACKSPACE through unformatted segmented files.
My code exists as a few separate programs, the first creates these very large, unformatted files, which are then read in by a second program.
I encounter the I/O performance drop at Intel compiler version 13 and onwards. The MPI code operates a simple read from the mesh file in the following way:
I am running into problems linking against acml 5.3.1 with ifort 14.0.1
while "gfortran -o TestACML TestACML.f90 -L /opt/acml5.3.1/gfortran64_int64/lib/" compiles without error,
"ifort -o TestACML TestACML.f90 -L/opt/acml5.3.1/ifort64_int64/lib/" produces: