mewin/glm
1
0
Fork 0
Code Releases Activity

Fixed sign with signed integer function on non-x86 architecture

Browse Source
This commit is contained in:
Christophe Riccio
2016-05-24 13:29:18 +02:00
parent 1208eb63f7
commit 72c741d8ea
5 changed files with 117 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
glm/detail/func_common.inl
View File

@@ -115,7 +115,7 @@ namespace detail
}
};
template <typename T, precision P, template <class, precision> class vecType, bool isFloat = true, bool isSigned = true>
template <typename T, precision P, template <class, precision> class vecType, bool isFloat = true>
struct compute_sign
{
GLM_FUNC_QUALIFIER static vecType<T, P> call(vecType<T, P> const & x)
@@ -124,17 +124,9 @@ namespace detail
}
};
# if GLM_ARCH == GLM_ARCH_X86
template <typename T, precision P, template <class, precision> class vecType>
struct compute_sign<T, P, vecType, false, false>
{
GLM_FUNC_QUALIFIER static vecType<T, P> call(vecType<T, P> const & x)
{
return vecType<T, P>(glm::greaterThan(x , vecType<T, P>(0)));
}
};
template <typename T, precision P, template <class, precision> class vecType>
struct compute_sign<T, P, vecType, false, true>
struct compute_sign<T, P, vecType, false>
{
GLM_FUNC_QUALIFIER static vecType<T, P> call(vecType<T, P> const & x)
{
@@ -144,6 +136,7 @@ namespace detail
return (x >> Shift) | y;
}
};
# endif
template <typename T, precision P, template <class, precision> class vecType, typename genType, bool isFloat = true>
struct compute_mod

88
glm/detail/type_vec4_simd.inl
View File

@@ -62,60 +62,114 @@ namespace detail
};
template <precision P>
struct compute_vec4_and<int, P>
struct compute_vec4_and<int32, P>
{
static tvec4<int, P> call(tvec4<int, P> const& a, tvec4<int, P> const& b)
static tvec4<int32, P> call(tvec4<int32, P> const& a, tvec4<int32, P> const& b)
{
tvec4<int, P> Result(uninitialize);
tvec4<int32, P> Result(uninitialize);
Result.data = _mm_and_si128(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_or<int, P>
struct compute_vec4_and<uint32, P>
{
static tvec4<int, P> call(tvec4<int, P> const& a, tvec4<int, P> const& b)
static tvec4<uint32, P> call(tvec4<uint32, P> const& a, tvec4<uint32, P> const& b)
{
tvec4<int, P> Result(uninitialize);
tvec4<uint32, P> Result(uninitialize);
Result.data = _mm_and_si128(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_or<int32, P>
{
static tvec4<int32, P> call(tvec4<int32, P> const& a, tvec4<int32, P> const& b)
{
tvec4<int32, P> Result(uninitialize);
Result.data = _mm_or_si128(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_xor<int, P>
struct compute_vec4_or<uint32, P>
{
static tvec4<int, P> call(tvec4<int, P> const& a, tvec4<int, P> const& b)
static tvec4<uint32, P> call(tvec4<uint32, P> const& a, tvec4<uint32, P> const& b)
{
tvec4<int, P> Result(uninitialize);
tvec4<uint32, P> Result(uninitialize);
Result.data = _mm_or_si128(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_xor<int32, P>
{
static tvec4<int32, P> call(tvec4<int32, P> const& a, tvec4<int32, P> const& b)
{
tvec4<int32, P> Result(uninitialize);
Result.data = _mm_xor_si128(a.data, b.data);
return Result;
}
};
/*
template <precision P>
struct compute_vec4_shift_left<int, P>
struct compute_vec4_xor<uint32, P>
{
static tvec4<int, P> call(tvec4<int, P> const& a, tvec4<int, P> const& b)
static tvec4<uint32, P> call(tvec4<uint32, P> const& a, tvec4<uint32, P> const& b)
{
tvec4<int, P> Result(uninitialize);
tvec4<uint32, P> Result(uninitialize);
Result.data = _mm_xor_si128(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_shift_left<int32, P>
{
static tvec4<int32, P> call(tvec4<int32, P> const& a, tvec4<int32, P> const& b)
{
tvec4<int32, P> Result(uninitialize);
Result.data = _mm_sll_epi32(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_shift_right<int, P>
struct compute_vec4_shift_left<uint32, P>
{
static tvec4<int, P> call(tvec4<int, P> const& a, tvec4<int, P> const& b)
static tvec4<uint32, P> call(tvec4<uint32, P> const& a, tvec4<uint32, P> const& b)
{
tvec4<int, P> Result(uninitialize);
tvec4<uint32, P> Result(uninitialize);
Result.data = _mm_sll_epi32(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_shift_right<int32, P>
{
static tvec4<int32, P> call(tvec4<int32, P> const& a, tvec4<int32, P> const& b)
{
tvec4<int32, P> Result(uninitialize);
Result.data = _mm_srl_epi32(a.data, b.data);
return Result;
}
};
template <precision P>
struct compute_vec4_shift_right<uint32, P>
{
static tvec4<uint32, P> call(tvec4<uint32, P> const& a, tvec4<uint32, P> const& b)
{
tvec4<uint32, P> Result(uninitialize);
Result.data = _mm_srl_epi32(a.data, b.data);
return Result;
}
};
*/
}//namespace detail
# if !GLM_HAS_DEFAULTED_FUNCTIONS