Cibiv · joshlvmh · Nov 4, 2021 · Nov 4, 2021 · Nov 4, 2021 · Nov 4, 2021
diff --git a/.gitignore b/.gitignore
@@ -22,3 +22,48 @@ build/
 /GTAGS
 /.dir-locals.el
 /iqtree_config.h
+eigen/*
+CPackConfig.cmake
+CPackSourceConfig.cmake
+Makefile
+alignment/CMakeFiles/
+alignment/Makefile
+alignment/cmake_install.cmake
+cmake_utils/neon_test_build/*
+example
+gsl/CmakeFiles/
+gsl/Makefile
+gsl/cmake_install.cmake
+lbfgsb/CMakeFiles/
+lbfgsb/Makefile
+lbfgsb/cmake_install.cmake
+model/CMakeFiles/
+model/cmake_install.cmake
+model/Makefile
+ncl/CMakeFiles/
+ncl/cmake_install.cmake
+ncl/Makefile
+nclextra/CMakeFiles/
+nclextra/cmake_install.cmake
+nclextra/Makefile
+neon_test_build
+pda/CMakeFiles/
+pda/cmake_install.cmake
+pda/Makefile
+sprng/CMakeFiles/
+sprng/cmake_install.cmake
+sprng/Makefile
+tree/CMakeFiles/
+tree/cmake_install.cmake
+tree/Makefile
+cmake_install.cmake
+utils/CMakeFiles/
+utils/cmake_install.cmake
+utils/Makefile
+vectorclass/CMakeFiles/
+vectorclass/cmake_install.cmake
+vectorclass/Makefile
+whtest/CMakeFiles/
+whtest/cmake_install.cmake
+whtest/Makefile
+includelog
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -76,6 +76,15 @@ endif()
 
 include_directories("${PROJECT_SOURCE_DIR}")
 
+# Check if __ARM_NEON is defined
+include(cmake_utls/check_neon_exists.cmake)
+if (NEON)
+set (__ARM_NEON "TRUE")
+message("__ARM_NEON DETECTED : TRUE")
+else()
+  message("__ARM_NEON DETECTED : FALSE")
+endif()
+
 ##################################################################
 # Detect target platforms
 ##################################################################
@@ -120,7 +129,11 @@ set(CLANG "FALSE") # Clang compiler
 set(ICC "FALSE")   # Intel compiler
 set(VCC "FALSE")   # MS Visual C Compiler, note that it is different from MSVC variable
 # using C++11 standard
-set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -Wno-deprecated")
+# disable AVX for NEON
+if (__ARM_NEON)
+set(IQTREE_FLAGS "${IQTREE_FLAGS} novx")
+endif()
 
 if (CMAKE_COMPILER_IS_GNUCXX)
 	message("Compiler      : GNU Compiler (gcc)")
@@ -306,9 +319,17 @@ SET(AVX_FLAGS "-D__SSE3 -D__AVX")
 if (VCC) 
 	set(AVX_FLAGS "${AVX_FLAGS} /arch:AVX")
 elseif (CLANG)
-	set(AVX_FLAGS "${AVX_FLAGS} -mavx")
+  if (__ARM_NEON)
+	set(AVX_FLAGS "${AVX_FLAGS} -march=armv8-a+fp+simd+crypto+crc")
+  else()
+        set(AVX_FLAGS "${AVX_FLAGS} -mavx")
+  endif()
 elseif (GCC)
-	set(AVX_FLAGS "${AVX_FLAGS} -mavx -fabi-version=0")
+  if (__ARM_NEON)
+	set(AVX_FLAGS "${AVX_FLAGS} -march=armv8-a+fp+simd+crypto+crc") #-mavx -fabi-version=0")
+  else()
+    set(AVX_FLAGS "${AVX_FLAGS} -mavx -fabi-version=0")
+  endif()
 elseif (ICC)
 	if (WIN32)
 		 set(AVX_FLAGS "${AVX_FLAGS} /arch:avx")
@@ -321,7 +342,11 @@ SET(SSE_FLAGS "-D__SSE3")
 if (VCC)
 	set(SSE_FLAGS "${SSE_FLAGS} /arch:SSE2 -D__SSE3__")
 elseif (GCC OR CLANG)
-	set(SSE_FLAGS "${SSE_FLAGS} -msse3")
+  if (__ARM_NEON)
+	set(SSE_FLAGS "${SSE_FLAGS} -march=armv8-a+fp+simd+crypto+crc")  # changed from -msse3
+  else()
+    set(SSE_FLAGS "${SSE_FLAGS} -msse3")
+  endif()
 elseif (ICC)
 	if (WIN32)
 		set(SSE_FLAGS "${SSE_FLAGS} /arch:sse3")
@@ -334,9 +359,17 @@ SET(FMA_FLAGS "-D__SSE3 -D__AVX")
 if (VCC) 
 	set(FMA_FLAGS "${FMA_FLAGS} /arch:AVX2")
 elseif (CLANG)
-	set(FMA_FLAGS "${FMA_FLAGS} -mavx -mfma")
+  if (__ARM_NEON)
+	set(FMA_FLAGS "${FMA_FLAGS} -march=armv8-a+fp+simd+crypto+crc") #-mavx -mfma")
+  else()
+    set(FMA_FLAGS "${FMA_FLAGS} -mavx -mfma")
+  endif()
 elseif (GCC)
-	set(FMA_FLAGS "${FMA_FLAGS} -mavx -fabi-version=0 -mfma")
+  if (__ARM_NEON)
+	set(FMA_FLAGS "${FMA_FLAGS} -march=armv8-a+fp+simd+crypto+crc") #-mavx -fabi-version=0 -mfma")
+  else()
+    set(FMA_FLAGS "${FMA_FLAGS} -mavx -fabi-version=0 -mfma")
+  endif()
 elseif (ICC) 
 	if (WIN32)
 		 set(FMA_FLAGS "${FMA_FLAGS} /arch:core-avx2")

diff --git a/cmake_utls/check_neon_exists.cmake b/cmake_utls/check_neon_exists.cmake
@@ -0,0 +1,17 @@
+cmake_minimum_required(VERSION 2.8.12)
+
+if (DEFINED NEON)
+  return()
+endif()
+
+set(NEON 0)
+
+enable_language(C)
+enable_language(CXX)
+
+try_compile(test_for_neon_worked ${PROJECT_BINARY_DIR}/neon_test_build ${CMAKE_CURRENT_LIST_DIR}/test_for_neon
+  neon_test)
+
+if(test_for_neon_worked)
+  set(NEON 1)
+endif()
diff --git a/cmake_utls/test_for_neon/CMakeLists.txt b/cmake_utls/test_for_neon/CMakeLists.txt
@@ -0,0 +1,4 @@
+cmake_minimum_required(VERSION 2.8.12)
+
+project(neon_test)
+add_library(neon_test STATIC neon_test.cpp)
diff --git a/cmake_utls/test_for_neon/neon_test.cpp b/cmake_utls/test_for_neon/neon_test.cpp
@@ -0,0 +1,5 @@
+#ifdef __ARM_NEON
+#else
+#error "No NEON"
+#endif
+int main(){}
diff --git a/example/example.cf b/example/example.cf
diff --git a/example/example.nex b/example/example.nex
diff --git a/example/example.phy b/example/example.phy
diff --git a/example/models.nex b/example/models.nex
diff --git a/intrinsics b/intrinsics
@@ -0,0 +1,33 @@
+_mm_andnot_pd’ was not declared in this scope;
+_mm_and_pd’ was not declared in this scope;
+_mm_blendv_pd’ was not declared in this scope;
+_mm_blendv_ps’ was not declared in this scope;
+_mm_castpd_ps’ was not declared in this scope;
+_mm_castpd_si128’ was not declared in this scope;
+_mm_castps_pd’ was not declared in this scope;
+_mm_castsi128_pd’ was not declared in this scope;
+_mm_cmpeq_pd’ was not declared in this scope;
+_mm_cmple_pd’ was not declared in this scope;
+_mm_cmplt_pd’ was not declared in this scope;
+_mm_cmpneq_pd’ was not declared in this scope;
+_mm_cvtepi32_pd’ was not declared in this scope;
+_mm_cvtpd_epi32’ was not declared in this scope;
+_mm_cvtpd_ps’ was not declared in this scope;
+_mm_cvtps_pd’ was not declared in this scope;
+_mm_cvtsd_f64’ was not declared in this scope;
+_mm_cvttpd_epi32’ was not declared in this scope;
+_mm_div_pd’ was not declared in this scope;
+_mm_getcsr’ was not declared in this scope;
+_mm_hadd_pd’ was not declared in this scope;
+_mm_movemask_pd’ was not declared in this scope;
+_mm_mulhi_epu16’ was not declared in this scope;
+_mm_mul_pd’ was not declared in this scope;
+_mm_or_pd’ was not declared in this scope;
+_mm_setcsr’ was not declared in this scope;
+_mm_setr_pd’ was not declared in this scope;
+_mm_set_sd’ was not declared in this scope;
+_mm_setzero_pd’ was not declared in this scope;
+_mm_shuffle_pd’ was not declared in this scope;
+_mm_store_sd’ was not declared in this scope;
+_mm_testc_si128’ was not declared in this scope;
+_mm_testz_si128’ was not declared in this scope;
diff --git a/lib/.!14422!libiomp5md.dll b/lib/.!14422!libiomp5md.dll
diff --git a/lib/.!14423!libomp.a b/lib/.!14423!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36032     `
diff --git a/lib/.!14424!libompstatic.a b/lib/.!14424!libompstatic.a
diff --git a/lib/.!14425!pthreadGC2.dll b/lib/.!14425!pthreadGC2.dll
diff --git a/lib/.!14426!pthreadGC2_64.dll b/lib/.!14426!pthreadGC2_64.dll
diff --git a/lib/.!14427!pthreadVC2.dll b/lib/.!14427!pthreadVC2.dll
diff --git a/lib/.!14428!pthreadVC2.lib b/lib/.!14428!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338093862              0       7120      `
diff --git a/lib/.!2826!libiomp5md.dll b/lib/.!2826!libiomp5md.dll
diff --git a/lib/.!2826!libiomp5md.dllr b/lib/.!2826!libiomp5md.dllr
diff --git a/lib/.!2827!libomp.a b/lib/.!2827!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36032     `
diff --git a/lib/.!2827!libomp.ar b/lib/.!2827!libomp.ar
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36032     `
diff --git a/lib/.!2828!libompstatic.a b/lib/.!2828!libompstatic.a
diff --git a/lib/.!2828!libompstatic.ar b/lib/.!2828!libompstatic.ar
diff --git a/lib/.!2829!pthreadGC2.dll b/lib/.!2829!pthreadGC2.dll
diff --git a/lib/.!2829!pthreadGC2.dllr b/lib/.!2829!pthreadGC2.dllr
diff --git a/lib/.!2830!pthreadGC2_64.dll b/lib/.!2830!pthreadGC2_64.dll
diff --git a/lib/.!2830!pthreadGC2_64.dllr b/lib/.!2830!pthreadGC2_64.dllr
diff --git a/lib/.!2831!pthreadVC2.dll b/lib/.!2831!pthreadVC2.dll
diff --git a/lib/.!2831!pthreadVC2.dllr b/lib/.!2831!pthreadVC2.dllr
diff --git a/lib/.!2832!pthreadVC2.lib b/lib/.!2832!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338093862              0       7120      `
diff --git a/lib/.!2832!pthreadVC2.libr b/lib/.!2832!pthreadVC2.libr
@@ -0,0 +1,2 @@
+!<arch>
+/               1338093862              0       7120      `
diff --git a/lib/.!6100!libiomp5md.dll b/lib/.!6100!libiomp5md.dll
diff --git a/lib/.!6101!libomp.a b/lib/.!6101!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36032     `
diff --git a/lib/.!6102!libompstatic.a b/lib/.!6102!libompstatic.a
diff --git a/lib/.!6103!pthreadGC2.dll b/lib/.!6103!pthreadGC2.dll
diff --git a/lib/.!6104!pthreadGC2_64.dll b/lib/.!6104!pthreadGC2_64.dll
diff --git a/lib/.!6105!pthreadVC2.dll b/lib/.!6105!pthreadVC2.dll
diff --git a/lib/.!6106!pthreadVC2.lib b/lib/.!6106!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338093862              0       7120      `
diff --git a/lib/.!7279!libiomp5md.dll b/lib/.!7279!libiomp5md.dll
diff --git a/lib/.!7280!libomp.a b/lib/.!7280!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36032     `
diff --git a/lib/.!7281!libompstatic.a b/lib/.!7281!libompstatic.a
diff --git a/lib/.!7282!pthreadGC2.dll b/lib/.!7282!pthreadGC2.dll
diff --git a/lib/.!7283!pthreadGC2_64.dll b/lib/.!7283!pthreadGC2_64.dll
diff --git a/lib/.!7284!pthreadVC2.dll b/lib/.!7284!pthreadVC2.dll
diff --git a/lib/.!7285!pthreadVC2.lib b/lib/.!7285!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338093862              0       7120      `
diff --git a/lib/.!9636!libiomp5md.dll b/lib/.!9636!libiomp5md.dll
diff --git a/lib/.!9637!libomp.a b/lib/.!9637!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36032     `
diff --git a/lib/.!9638!libompstatic.a b/lib/.!9638!libompstatic.a
diff --git a/lib/.!9639!pthreadGC2.dll b/lib/.!9639!pthreadGC2.dll
diff --git a/lib/.!9640!pthreadGC2_64.dll b/lib/.!9640!pthreadGC2_64.dll
diff --git a/lib/.!9641!pthreadVC2.dll b/lib/.!9641!pthreadVC2.dll
diff --git a/lib/.!9642!pthreadVC2.lib b/lib/.!9642!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338093862              0       7120      `
diff --git a/lib32/.!14473!libiomp5md.dll b/lib32/.!14473!libiomp5md.dll
diff --git a/lib32/.!14474!libiomp5md.lib.dll b/lib32/.!14474!libiomp5md.lib.dll
diff --git a/lib32/.!14475!libomp.a b/lib32/.!14475!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36018     `
diff --git a/lib32/.!14476!pthreadVC2.dll b/lib32/.!14476!pthreadVC2.dll
diff --git a/lib32/.!14477!pthreadVC2.lib b/lib32/.!14477!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338046565              0       7360      `
diff --git a/lib32/.!2833!libiomp5md.dll b/lib32/.!2833!libiomp5md.dll
diff --git a/lib32/.!2833!libiomp5md.dllr b/lib32/.!2833!libiomp5md.dllr
diff --git a/lib32/.!2834!libiomp5md.lib.dll b/lib32/.!2834!libiomp5md.lib.dll
diff --git a/lib32/.!2834!libiomp5md.lib.dllr b/lib32/.!2834!libiomp5md.lib.dllr
diff --git a/lib32/.!2835!libomp.a b/lib32/.!2835!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36018     `
diff --git a/lib32/.!2835!libomp.ar b/lib32/.!2835!libomp.ar
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36018     `
diff --git a/lib32/.!2836!pthreadVC2.dll b/lib32/.!2836!pthreadVC2.dll
diff --git a/lib32/.!2836!pthreadVC2.dllr b/lib32/.!2836!pthreadVC2.dllr
diff --git a/lib32/.!2837!pthreadVC2.lib b/lib32/.!2837!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338046565              0       7360      `
diff --git a/lib32/.!2837!pthreadVC2.libr b/lib32/.!2837!pthreadVC2.libr
@@ -0,0 +1,2 @@
+!<arch>
+/               1338046565              0       7360      `
diff --git a/lib32/.!6112!libiomp5md.dll b/lib32/.!6112!libiomp5md.dll
diff --git a/lib32/.!6113!libiomp5md.lib.dll b/lib32/.!6113!libiomp5md.lib.dll
diff --git a/lib32/.!6114!libomp.a b/lib32/.!6114!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36018     `
diff --git a/lib32/.!6115!pthreadVC2.dll b/lib32/.!6115!pthreadVC2.dll
diff --git a/lib32/.!6116!pthreadVC2.lib b/lib32/.!6116!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338046565              0       7360      `
diff --git a/lib32/.!7301!libiomp5md.dll b/lib32/.!7301!libiomp5md.dll
diff --git a/lib32/.!7302!libiomp5md.lib.dll b/lib32/.!7302!libiomp5md.lib.dll
diff --git a/lib32/.!7303!libomp.a b/lib32/.!7303!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+/               0           0     0     0       36018     `
diff --git a/lib32/.!7304!pthreadVC2.dll b/lib32/.!7304!pthreadVC2.dll
diff --git a/lib32/.!7305!pthreadVC2.lib b/lib32/.!7305!pthreadVC2.lib
@@ -0,0 +1,2 @@
+!<arch>
+/               1338046565              0       7360      `
diff --git a/libmac/.!14485!libomp.a b/libmac/.!14485!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+#1/20           1530061535  502   20    100644  39820     `
diff --git a/libmac/.!2838!libomp.a b/libmac/.!2838!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+#1/20           1530061535  502   20    100644  39820     `
diff --git a/libmac/.!2838!libomp.ar b/libmac/.!2838!libomp.ar
@@ -0,0 +1,2 @@
+!<arch>
+#1/20           1530061535  502   20    100644  39820     `
diff --git a/libmac/.!6118!libomp.a b/libmac/.!6118!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+#1/20           1530061535  502   20    100644  39820     `
diff --git a/libmac/.!7309!libomp.a b/libmac/.!7309!libomp.a
@@ -0,0 +1,2 @@
+!<arch>
+#1/20           1530061535  502   20    100644  39820     `
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,2 @@
		!<arch>
		#1/20 1530061535 502 20 100644 39820 `