64-BitBrainstorm_2026/AMSS-NCKU
8
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refactor verification method and optimize numerical kernels with oneMKL BLAS

Browse Source 
This commit transitions the verification approach from post-Newtonian theory
   comparison to regression testing against baseline simulations, and optimizes
   critical numerical kernels using Intel oneMKL BLAS routines.

   Verification Changes:
   - Replace PN theory-based RMS calculation with trajectory-based comparison
   - Compare optimized results against baseline (GW150914-origin) on XY plane
   - Compute RMS independently for BH1 and BH2, report maximum as final metric
   - Update documentation to reflect new regression test methodology

   Performance Optimizations:
   - Replace manual vector operations with oneMKL BLAS routines:
     * norm2() and scalarproduct() now use cblas_dnrm2/cblas_ddot (C++)
     * L2 norm calculations use DDOT for dot products (Fortran)
     * Interpolation weighted sums use DDOT (Fortran)
   - Disable OpenMP threading (switch to sequential MKL) for better performance

   Build Configuration:
   - Switch from lmkl_intel_thread to lmkl_sequential
   - Remove -qopenmp flags from compiler options
   - Maintain aggressive optimization flags (-O3, -xHost, -fp-model fast=2, -fma)

   Other Changes:
   - Update .gitignore for GW150914-origin, docs, and temporary files
This commit is contained in:
CGH0S7
2026-01-18 14:25:21 +08:00
parent 3a7bce3af2
commit 9deeda9831
5 changed files with 170 additions and 99 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
AMSS_NCKU_source/makefile.inc
View File

@@ -6,20 +6,20 @@
## Intel oneAPI version with oneMKL (Optimized for performance)
filein = -I/usr/include/ -I${MKLROOT}/include
## Use Intel OpenMP threading layer for better performance
## Using sequential MKL (OpenMP disabled for better single-threaded performance)
LDLIBS = -L/usr/lib/x86_64-linux-gnu -L/usr/lib64 -lifcore -limf -lmpi \
-L${MKLROOT}/lib -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core \
-liomp5 -lpthread -lm -ldl
-L${MKLROOT}/lib -lmkl_intel_lp64 -lmkl_sequential -lmkl_core \
-lpthread -lm -ldl
## Aggressive optimization flags:
## -O3: Maximum optimization
## -xHost: Optimize for the host CPU architecture (Intel/AMD compatible)
## -qopenmp: Enable OpenMP parallelization
## -fp-model fast=2: Aggressive floating-point optimizations
## -fma: Enable fused multiply-add instructions
CXXAPPFLAGS = -O3 -xHost -qopenmp -fp-model fast=2 -fma \
## Note: OpenMP has been disabled (-qopenmp removed) due to performance issues
CXXAPPFLAGS = -O3 -xHost -fp-model fast=2 -fma \
-Dfortran3 -Dnewc -I${MKLROOT}/include
f90appflags = -O3 -xHost -qopenmp -fp-model fast=2 -fma \
f90appflags = -O3 -xHost -fp-model fast=2 -fma \
-fpp -I${MKLROOT}/include
f90 = ifx
f77 = ifx