mirror/glm
1
0
Fork 0
mirror of https://github.com/g-truc/glm.git synced 2024-11-30 03:44:38 +00:00
Code Issues Packages Projects Releases Wiki Activity

Revert broken vec4 SIMD

Browse Source
This commit is contained in:
Christophe 2023-12-20 21:55:32 +01:00
parent 8bdcf22d72
commit 857257745d
1 changed files with 425 additions and 447 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
glm/detail/type_vec4_simd.inl
View File

@ -1,11 +1,11 @@
#if GLM_ARCH & GLM_ARCH_SSE2_BIT #if GLM_ARCH & GLM_ARCH_SSE2_BIT
namespace glm{ namespace glm {
namespace detail namespace detail
{ {
# if GLM_CONFIG_SWIZZLE == GLM_SWIZZLE_OPERATOR # if GLM_CONFIG_SWIZZLE == GLM_SWIZZLE_OPERATOR
template<qualifier Q, int E0, int E1, int E2, int E3> template<qualifier Q, int E0, int E1, int E2, int E3>
struct _swizzle_base1<4, float, Q, E0,E1,E2,E3, true> : public _swizzle_base0<float, 4> struct _swizzle_base1<4, float, Q, E0, E1, E2, E3, true> : public _swizzle_base0<float, 4>
{ {
GLM_FUNC_QUALIFIER vec<4, float, Q> operator ()() const GLM_FUNC_QUALIFIER vec<4, float, Q> operator ()() const
{ {
@ -22,7 +22,7 @@ namespace detail
}; };
template<qualifier Q, int E0, int E1, int E2, int E3> template<qualifier Q, int E0, int E1, int E2, int E3>
struct _swizzle_base1<4, int, Q, E0,E1,E2,E3, true> : public _swizzle_base0<int, 4> struct _swizzle_base1<4, int, Q, E0, E1, E2, E3, true> : public _swizzle_base0<int, 4>
{ {
GLM_FUNC_QUALIFIER vec<4, int, Q> operator ()() const GLM_FUNC_QUALIFIER vec<4, int, Q> operator ()() const
{ {
@ -35,7 +35,7 @@ namespace detail
}; };
template<qualifier Q, int E0, int E1, int E2, int E3> template<qualifier Q, int E0, int E1, int E2, int E3>
struct _swizzle_base1<4, uint, Q, E0,E1,E2,E3, true> : public _swizzle_base0<uint, 4> struct _swizzle_base1<4, uint, Q, E0, E1, E2, E3, true> : public _swizzle_base0<uint, 4>
{ {
GLM_FUNC_QUALIFIER vec<4, uint, Q> operator ()() const GLM_FUNC_QUALIFIER vec<4, uint, Q> operator ()() const
{ {
@ -156,7 +156,7 @@ namespace detail
}; };
template<typename T, qualifier Q> template<typename T, qualifier Q>
struct compute_vec4_and<uint, Q, true, 32, true> struct compute_vec4_and<T, Q, true, 32, true>
{ {
static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b) static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b)
{ {
@ -180,7 +180,7 @@ namespace detail
# endif # endif
template<typename T, qualifier Q> template<typename T, qualifier Q>
struct compute_vec4_or<int, Q, true, 32, true> struct compute_vec4_or<T, Q, true, 32, true>
{ {
static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b) static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b)
{ {
@ -191,19 +191,8 @@ namespace detail
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_BIT # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template<qualifier Q>
struct compute_vec4_or<int64, Q, true, 64, true>
{
GLM_FUNC_QUALIFIER static vec<4, int64, Q> call(vec<4, int64, Q> const& a, vec<4, int64, Q> const& b)
{
vec<4, int64, Q> Result;
Result.data = _mm256_or_si256(a.data, b.data);
return Result;
}
};
template<typename T, qualifier Q> template<typename T, qualifier Q>
struct compute_vec4_or<uint64, Q, true, 64, true> struct compute_vec4_or<T, Q, true, 64, true>
{ {
static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b) static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b)
{ {
@ -215,7 +204,7 @@ namespace detail
# endif # endif
template<typename T, qualifier Q> template<typename T, qualifier Q>
struct compute_vec4_xor<int, Q, true, 32, true> struct compute_vec4_xor<T, Q, true, 32, true>
{ {
static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b) static vec<4, T, Q> call(vec<4, T, Q> const& a, vec<4, T, Q> const& b)
{ {
@ -274,17 +263,6 @@ namespace detail
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_BIT # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template<typename T, qualifier Q>
struct compute_vec4_shift_right<int64, Q, true, 64, true>
{
GLM_FUNC_QUALIFIER static vec<4, int64, Q> call(vec<4, int64, Q> const& a, vec<4, int64, Q> const& b)
{
vec<4, int64, Q> Result;
Result.data = _mm256_srl_epi64(a.data, b.data);
return Result;
}
};
template<typename T, qualifier Q> template<typename T, qualifier Q>
struct compute_vec4_shift_right<T, Q, true, 64, true> struct compute_vec4_shift_right<T, Q, true, 64, true>
{ {
@ -364,7 +342,7 @@ namespace detail
} }
}; };
# endif # endif
}//namespace detail }//namespace detail
template<> template<>
GLM_FUNC_QUALIFIER GLM_CONSTEXPR vec<4, float, aligned_lowp>::vec(float _s) : GLM_FUNC_QUALIFIER GLM_CONSTEXPR vec<4, float, aligned_lowp>::vec(float _s) :
@ -486,7 +464,7 @@ namespace detail
#if GLM_ARCH & GLM_ARCH_NEON_BIT #if GLM_ARCH & GLM_ARCH_NEON_BIT
namespace glm { namespace glm {
namespace detail { namespace detail {
template<qualifier Q> template<qualifier Q>
struct compute_vec4_add<float, Q, true> struct compute_vec4_add<float, Q, true>
@ -701,7 +679,7 @@ namespace detail {
} }
}; };
}//namespace detail }//namespace detail
#if !GLM_CONFIG_XYZW_ONLY #if !GLM_CONFIG_XYZW_ONLY
template<> template<>