Add find first set for simd masks to SIMDWrapper

Kernel/LHCbMath/include/LHCbMath/SIMDWrapper.h

@@ -220,6 +220,7 @@ namespace SIMDWrapper {
       constexpr friend bool none( mask_v mask ) { return mask.data == 0; }
       constexpr friend bool any( mask_v mask ) { return mask.data != 0; }
       constexpr friend bool testbit( mask_v mask, int ) { return mask.data == 1; }
+      constexpr friend int  ffs( mask_v mask ) { return mask.data; }
 
       friend std::ostream& operator<<( std::ostream& os, mask_v mask ) { return os << "scalar{" << mask.data << "}"; }
 
@@ -520,6 +521,8 @@ namespace SIMDWrapper {
       friend bool none( mask_v mask ) { return movemask_u32( mask.native() ) == 0x0; }
       friend bool any( mask_v mask ) { return movemask_u32( mask.native() ) != 0x0; }
       friend bool testbit( mask_v mask, int bit ) { return ( movemask_u32( mask.native() ) & ( 1 << bit ) ) != 0; }
+      // TODO: c++20 countl_zero
+      friend int ffs( mask_v mask ) { return __builtin_ffs( movemask_u32( mask.native() ) ); }
 
       friend std::ostream& operator<<( std::ostream& os, mask_v x ) {
         return print_bitset<4>( os, static_cast<unsigned long long>( movemask_u32( x.native() ) ), "neon" );
@@ -902,6 +905,8 @@ namespace SIMDWrapper {
       friend bool none( mask_v mask ) { return _mm_movemask_ps( mask.native() ) == 0x0; }
       friend bool any( mask_v mask ) { return _mm_movemask_ps( mask.native() ) != 0x0; }
       friend bool testbit( mask_v mask, int bit ) { return ( _mm_movemask_ps( mask.native() ) & ( 1 << bit ) ) != 0; }
+      // TODO: c++20 countl_zero
+      friend int ffs( mask_v mask ) { return __builtin_ffs( _mm_movemask_ps( mask.native() ) ); }
 
       friend std::ostream& operator<<( std::ostream& os, mask_v x ) {
         return print_bitset<4>( os, static_cast<unsigned long long>( _mm_movemask_ps( x.native() ) ), "sse" );
@@ -1345,6 +1350,8 @@ namespace SIMDWrapper {
       friend bool testbit( mask_v mask, const int bit ) {
         return ( _mm256_movemask_ps( mask.native() ) & ( 1 << bit ) ) != 0;
       }
+      // TODO: c++20 countl_zero
+      friend int ffs( mask_v mask ) { return __builtin_ffs( _mm256_movemask_ps( mask.native() ) ); }
 
       __m256 native() const { return data; }
 
@@ -1694,6 +1701,8 @@ namespace SIMDWrapper {
       friend bool none( mask_v mask ) { return mask.native() == 0x00; }
       friend bool any( mask_v mask ) { return mask.native() != 0x00; }
       friend bool testbit( mask_v mask, int bit ) { return ( mask.native() & ( 1 << bit ) ) != 0; }
+      // TODO: c++20 countl_zero
+      friend int ffs( mask_v mask ) { return __builtin_ffs( mask.native() ); }
 
       friend std::ostream& operator<<( std::ostream& os, mask_v x ) {
         return print_bitset<8>( os, x.native(), "avx256" );
@@ -2013,6 +2022,8 @@ namespace SIMDWrapper {
       friend bool none( mask_v mask ) { return mask.native() == 0x0000; }
       friend bool any( mask_v mask ) { return mask.native() != 0x0000; }
       friend bool testbit( mask_v mask, int bit ) { return ( mask.native() & ( 1 << bit ) ) != 0; }
+      // TODO: c++20 countl_zero
+      friend int ffs( mask_v mask ) { return __builtin_ffs( mask.native() ); }
 
       friend std::ostream& operator<<( std::ostream& os, mask_v x ) {
         return print_bitset<16>( os, x.native(), "avx512" );