zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
gatk-3.8/public/c++/VectorPairHMM/avx_function_instantiations.cc

20 lines
458 B
C++
Raw Normal View History

Added support for dynamic selection between AVX and un-vectorized C++, still to include SSE code from Mohammad. Debug flags turned on in this commit.
2014-01-19 03:07:23 +08:00
#include "template.h"
Clean up
2014-01-27 03:36:06 +08:00
#undef SIMD_ENGINE
#undef SIMD_ENGINE_SSE
1. Integrated Mohammad's SSE4.2 code, Mustafa's bug fix and code to fix the SSE compilation warning. 2. Added code to dynamically select between AVX, SSE4.2 and normal C++ (in that order) 3. Created multiple files to compile with different compilation flags: avx_function_prototypes.cc is compiled with -xAVX while sse_function_instantiations.cc is compiled with -xSSE4.2 flag. 4. Added jniClose() and support in Java (HaplotypeCaller, PairHMMLikelihoodCalculationEngine) to call this function at the end of the program. 5. Removed debug code, kept assertions and profiling in C++ 6. Disabled OpenMP for now.
2014-01-21 00:03:42 +08:00
Clean up
2014-01-27 03:36:06 +08:00
#define SIMD_ENGINE avx
#define SIMD_ENGINE_AVX
1. Integrated Mohammad's SSE4.2 code, Mustafa's bug fix and code to fix the SSE compilation warning. 2. Added code to dynamically select between AVX, SSE4.2 and normal C++ (in that order) 3. Created multiple files to compile with different compilation flags: avx_function_prototypes.cc is compiled with -xAVX while sse_function_instantiations.cc is compiled with -xSSE4.2 flag. 4. Added jniClose() and support in Java (HaplotypeCaller, PairHMMLikelihoodCalculationEngine) to call this function at the end of the program. 5. Removed debug code, kept assertions and profiling in C++ 6. Disabled OpenMP for now.
2014-01-21 00:03:42 +08:00
Added support for dynamic selection between AVX and un-vectorized C++, still to include SSE code from Mohammad. Debug flags turned on in this commit.
2014-01-19 03:07:23 +08:00
#include "define-float.h"
#include "shift_template.c"
#include "pairhmm-template-kernel.cc"
#include "define-double.h"
#include "shift_template.c"
#include "pairhmm-template-kernel.cc"
template double compute_full_prob_avxd<double>(testcase* tc, double* nextlog);
template float compute_full_prob_avxs<float>(testcase* tc, float* nextlog);
Got rid of the MMX instructions in the SSE version of the code. Handling the mask operations in a class, which is defined for each version of SSE and AVX implementations separately.
2014-01-22 01:30:15 +08:00