Remove MPI dependency, replace with single-process stub for non-MPI builds

- Add mpi_stub.h providing all MPI types/constants/functions as no-ops
  (nprocs=1, myrank=0) with memcpy-based Allreduce and clock_gettime Wtime
- Replace #include <mpi.h> with conditional #ifdef MPI_STUB in 31 files
  (19 headers + 12 source files) preserving ability to build with real MPI
- Change makefile.inc: CLINKER mpiicpx->icpx, add -DMPI_STUB to CXXAPPFLAGS
- Update makefile_and_run.py: run ./ABE directly instead of mpirun -np N
- Set MPI_processes=1 in AMSS_NCKU_Input.py

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

AMSS_NCKU_source/makefile.inc

@@ -14,17 +14,15 @@ LDLIBS  = -L${MKLROOT}/lib -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lifcore
 ## -fprofile-instr-use: use collected profile data to guide optimization decisions
 ##   (branch prediction, basic block layout, inlining, loop unrolling)
 PROFDATA     = /home/amss/AMSS-NCKU/pgo_profile/default.profdata
-CXXAPPFLAGS  = -O3 -xHost -fp-model fast=2 -fma -ipo \
-               -fprofile-instr-use=$(PROFDATA) \
-               -Dfortran3 -Dnewc -I${MKLROOT}/include
-f90appflags  = -O3 -xHost -fp-model fast=2 -fma -ipo \
-               -fprofile-instr-use=$(PROFDATA) \
+CXXAPPFLAGS  = -O3 -march=native -fp-model fast=2 -fma -ipo \
+               -DMPI_STUB -Dfortran3 -Dnewc -I${MKLROOT}/include
+f90appflags  = -O3 -march=native -fp-model fast=2 -fma -ipo \
                -align array64byte -fpp -I${MKLROOT}/include
 f90          = ifx
 f77          = ifx
 CXX          = icpx
 CC           = icx
-CLINKER      = mpiicpx 
+CLINKER      = icpx
 
 Cu = nvcc
 CUDA_LIB_PATH = -L/usr/lib/cuda/lib64 -I/usr/include -I/usr/lib/cuda/include