initial commit
This commit is contained in:
izenynn
240
lib/glm/simd/common.h
Normal file
240
lib/glm/simd/common.h
Normal file
@@ -0,0 +1,240 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/common.h
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "platform.h"
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_add(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_add_ps(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_add(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_add_ss(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_sub(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_sub_ps(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_sub(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_sub_ss(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_mul(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_mul_ps(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_mul(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_mul_ss(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_div(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_div_ps(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_div(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return _mm_div_ss(a, b);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_div_lowp(glm_vec4 a, glm_vec4 b)
|
||||
{
|
||||
return glm_vec4_mul(a, _mm_rcp_ps(b));
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_swizzle_xyzw(glm_vec4 a)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_AVX2_BIT
|
||||
return _mm_permute_ps(a, _MM_SHUFFLE(3, 2, 1, 0));
|
||||
# else
|
||||
return _mm_shuffle_ps(a, a, _MM_SHUFFLE(3, 2, 1, 0));
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_fma(glm_vec4 a, glm_vec4 b, glm_vec4 c)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_AVX2_BIT
|
||||
return _mm_fmadd_ss(a, b, c);
|
||||
# else
|
||||
return _mm_add_ss(_mm_mul_ss(a, b), c);
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_fma(glm_vec4 a, glm_vec4 b, glm_vec4 c)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_AVX2_BIT
|
||||
return _mm_fmadd_ps(a, b, c);
|
||||
# else
|
||||
return glm_vec4_add(glm_vec4_mul(a, b), c);
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_abs(glm_vec4 x)
|
||||
{
|
||||
return _mm_and_ps(x, _mm_castsi128_ps(_mm_set1_epi32(0x7FFFFFFF)));
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_ivec4 glm_ivec4_abs(glm_ivec4 x)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_SSSE3_BIT
|
||||
return _mm_sign_epi32(x, x);
|
||||
# else
|
||||
glm_ivec4 const sgn0 = _mm_srai_epi32(x, 31);
|
||||
glm_ivec4 const inv0 = _mm_xor_si128(x, sgn0);
|
||||
glm_ivec4 const sub0 = _mm_sub_epi32(inv0, sgn0);
|
||||
return sub0;
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_sign(glm_vec4 x)
|
||||
{
|
||||
glm_vec4 const zro0 = _mm_setzero_ps();
|
||||
glm_vec4 const cmp0 = _mm_cmplt_ps(x, zro0);
|
||||
glm_vec4 const cmp1 = _mm_cmpgt_ps(x, zro0);
|
||||
glm_vec4 const and0 = _mm_and_ps(cmp0, _mm_set1_ps(-1.0f));
|
||||
glm_vec4 const and1 = _mm_and_ps(cmp1, _mm_set1_ps(1.0f));
|
||||
glm_vec4 const or0 = _mm_or_ps(and0, and1);;
|
||||
return or0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_round(glm_vec4 x)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_SSE41_BIT
|
||||
return _mm_round_ps(x, _MM_FROUND_TO_NEAREST_INT);
|
||||
# else
|
||||
glm_vec4 const sgn0 = _mm_castsi128_ps(_mm_set1_epi32(0x80000000));
|
||||
glm_vec4 const and0 = _mm_and_ps(sgn0, x);
|
||||
glm_vec4 const or0 = _mm_or_ps(and0, _mm_set_ps1(8388608.0f));
|
||||
glm_vec4 const add0 = glm_vec4_add(x, or0);
|
||||
glm_vec4 const sub0 = glm_vec4_sub(add0, or0);
|
||||
return sub0;
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_floor(glm_vec4 x)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_SSE41_BIT
|
||||
return _mm_floor_ps(x);
|
||||
# else
|
||||
glm_vec4 const rnd0 = glm_vec4_round(x);
|
||||
glm_vec4 const cmp0 = _mm_cmplt_ps(x, rnd0);
|
||||
glm_vec4 const and0 = _mm_and_ps(cmp0, _mm_set1_ps(1.0f));
|
||||
glm_vec4 const sub0 = glm_vec4_sub(rnd0, and0);
|
||||
return sub0;
|
||||
# endif
|
||||
}
|
||||
|
||||
/* trunc TODO
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_trunc(glm_vec4 x)
|
||||
{
|
||||
return glm_vec4();
|
||||
}
|
||||
*/
|
||||
|
||||
//roundEven
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_roundEven(glm_vec4 x)
|
||||
{
|
||||
glm_vec4 const sgn0 = _mm_castsi128_ps(_mm_set1_epi32(0x80000000));
|
||||
glm_vec4 const and0 = _mm_and_ps(sgn0, x);
|
||||
glm_vec4 const or0 = _mm_or_ps(and0, _mm_set_ps1(8388608.0f));
|
||||
glm_vec4 const add0 = glm_vec4_add(x, or0);
|
||||
glm_vec4 const sub0 = glm_vec4_sub(add0, or0);
|
||||
return sub0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_ceil(glm_vec4 x)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_SSE41_BIT
|
||||
return _mm_ceil_ps(x);
|
||||
# else
|
||||
glm_vec4 const rnd0 = glm_vec4_round(x);
|
||||
glm_vec4 const cmp0 = _mm_cmpgt_ps(x, rnd0);
|
||||
glm_vec4 const and0 = _mm_and_ps(cmp0, _mm_set1_ps(1.0f));
|
||||
glm_vec4 const add0 = glm_vec4_add(rnd0, and0);
|
||||
return add0;
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_fract(glm_vec4 x)
|
||||
{
|
||||
glm_vec4 const flr0 = glm_vec4_floor(x);
|
||||
glm_vec4 const sub0 = glm_vec4_sub(x, flr0);
|
||||
return sub0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_mod(glm_vec4 x, glm_vec4 y)
|
||||
{
|
||||
glm_vec4 const div0 = glm_vec4_div(x, y);
|
||||
glm_vec4 const flr0 = glm_vec4_floor(div0);
|
||||
glm_vec4 const mul0 = glm_vec4_mul(y, flr0);
|
||||
glm_vec4 const sub0 = glm_vec4_sub(x, mul0);
|
||||
return sub0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_clamp(glm_vec4 v, glm_vec4 minVal, glm_vec4 maxVal)
|
||||
{
|
||||
glm_vec4 const min0 = _mm_min_ps(v, maxVal);
|
||||
glm_vec4 const max0 = _mm_max_ps(min0, minVal);
|
||||
return max0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_mix(glm_vec4 v1, glm_vec4 v2, glm_vec4 a)
|
||||
{
|
||||
glm_vec4 const sub0 = glm_vec4_sub(_mm_set1_ps(1.0f), a);
|
||||
glm_vec4 const mul0 = glm_vec4_mul(v1, sub0);
|
||||
glm_vec4 const mad0 = glm_vec4_fma(v2, a, mul0);
|
||||
return mad0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_step(glm_vec4 edge, glm_vec4 x)
|
||||
{
|
||||
glm_vec4 const cmp = _mm_cmple_ps(x, edge);
|
||||
return _mm_movemask_ps(cmp) == 0 ? _mm_set1_ps(1.0f) : _mm_setzero_ps();
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_smoothstep(glm_vec4 edge0, glm_vec4 edge1, glm_vec4 x)
|
||||
{
|
||||
glm_vec4 const sub0 = glm_vec4_sub(x, edge0);
|
||||
glm_vec4 const sub1 = glm_vec4_sub(edge1, edge0);
|
||||
glm_vec4 const div0 = glm_vec4_sub(sub0, sub1);
|
||||
glm_vec4 const clp0 = glm_vec4_clamp(div0, _mm_setzero_ps(), _mm_set1_ps(1.0f));
|
||||
glm_vec4 const mul0 = glm_vec4_mul(_mm_set1_ps(2.0f), clp0);
|
||||
glm_vec4 const sub2 = glm_vec4_sub(_mm_set1_ps(3.0f), mul0);
|
||||
glm_vec4 const mul1 = glm_vec4_mul(clp0, clp0);
|
||||
glm_vec4 const mul2 = glm_vec4_mul(mul1, sub2);
|
||||
return mul2;
|
||||
}
|
||||
|
||||
// Agner Fog method
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_nan(glm_vec4 x)
|
||||
{
|
||||
glm_ivec4 const t1 = _mm_castps_si128(x); // reinterpret as 32-bit integer
|
||||
glm_ivec4 const t2 = _mm_sll_epi32(t1, _mm_cvtsi32_si128(1)); // shift out sign bit
|
||||
glm_ivec4 const t3 = _mm_set1_epi32(0xFF000000); // exponent mask
|
||||
glm_ivec4 const t4 = _mm_and_si128(t2, t3); // exponent
|
||||
glm_ivec4 const t5 = _mm_andnot_si128(t3, t2); // fraction
|
||||
glm_ivec4 const Equal = _mm_cmpeq_epi32(t3, t4);
|
||||
glm_ivec4 const Nequal = _mm_cmpeq_epi32(t5, _mm_setzero_si128());
|
||||
glm_ivec4 const And = _mm_and_si128(Equal, Nequal);
|
||||
return _mm_castsi128_ps(And); // exponent = all 1s and fraction != 0
|
||||
}
|
||||
|
||||
// Agner Fog method
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_inf(glm_vec4 x)
|
||||
{
|
||||
glm_ivec4 const t1 = _mm_castps_si128(x); // reinterpret as 32-bit integer
|
||||
glm_ivec4 const t2 = _mm_sll_epi32(t1, _mm_cvtsi32_si128(1)); // shift out sign bit
|
||||
return _mm_castsi128_ps(_mm_cmpeq_epi32(t2, _mm_set1_epi32(0xFF000000))); // exponent is all 1s, fraction is 0
|
||||
}
|
||||
|
||||
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
20
lib/glm/simd/exponential.h
Normal file
20
lib/glm/simd/exponential.h
Normal file
@@ -0,0 +1,20 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/experimental.h
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "platform.h"
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_sqrt_lowp(glm_vec4 x)
|
||||
{
|
||||
return _mm_mul_ss(_mm_rsqrt_ss(x), x);
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_sqrt_lowp(glm_vec4 x)
|
||||
{
|
||||
return _mm_mul_ps(_mm_rsqrt_ps(x), x);
|
||||
}
|
||||
|
||||
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
124
lib/glm/simd/geometric.h
Normal file
124
lib/glm/simd/geometric.h
Normal file
@@ -0,0 +1,124 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/geometric.h
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "common.h"
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
GLM_FUNC_DECL glm_vec4 glm_vec4_dot(glm_vec4 v1, glm_vec4 v2);
|
||||
GLM_FUNC_DECL glm_vec4 glm_vec1_dot(glm_vec4 v1, glm_vec4 v2);
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_length(glm_vec4 x)
|
||||
{
|
||||
glm_vec4 const dot0 = glm_vec4_dot(x, x);
|
||||
glm_vec4 const sqt0 = _mm_sqrt_ps(dot0);
|
||||
return sqt0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_distance(glm_vec4 p0, glm_vec4 p1)
|
||||
{
|
||||
glm_vec4 const sub0 = _mm_sub_ps(p0, p1);
|
||||
glm_vec4 const len0 = glm_vec4_length(sub0);
|
||||
return len0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_dot(glm_vec4 v1, glm_vec4 v2)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_AVX_BIT
|
||||
return _mm_dp_ps(v1, v2, 0xff);
|
||||
# elif GLM_ARCH & GLM_ARCH_SSE3_BIT
|
||||
glm_vec4 const mul0 = _mm_mul_ps(v1, v2);
|
||||
glm_vec4 const hadd0 = _mm_hadd_ps(mul0, mul0);
|
||||
glm_vec4 const hadd1 = _mm_hadd_ps(hadd0, hadd0);
|
||||
return hadd1;
|
||||
# else
|
||||
glm_vec4 const mul0 = _mm_mul_ps(v1, v2);
|
||||
glm_vec4 const swp0 = _mm_shuffle_ps(mul0, mul0, _MM_SHUFFLE(2, 3, 0, 1));
|
||||
glm_vec4 const add0 = _mm_add_ps(mul0, swp0);
|
||||
glm_vec4 const swp1 = _mm_shuffle_ps(add0, add0, _MM_SHUFFLE(0, 1, 2, 3));
|
||||
glm_vec4 const add1 = _mm_add_ps(add0, swp1);
|
||||
return add1;
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_dot(glm_vec4 v1, glm_vec4 v2)
|
||||
{
|
||||
# if GLM_ARCH & GLM_ARCH_AVX_BIT
|
||||
return _mm_dp_ps(v1, v2, 0xff);
|
||||
# elif GLM_ARCH & GLM_ARCH_SSE3_BIT
|
||||
glm_vec4 const mul0 = _mm_mul_ps(v1, v2);
|
||||
glm_vec4 const had0 = _mm_hadd_ps(mul0, mul0);
|
||||
glm_vec4 const had1 = _mm_hadd_ps(had0, had0);
|
||||
return had1;
|
||||
# else
|
||||
glm_vec4 const mul0 = _mm_mul_ps(v1, v2);
|
||||
glm_vec4 const mov0 = _mm_movehl_ps(mul0, mul0);
|
||||
glm_vec4 const add0 = _mm_add_ps(mov0, mul0);
|
||||
glm_vec4 const swp1 = _mm_shuffle_ps(add0, add0, 1);
|
||||
glm_vec4 const add1 = _mm_add_ss(add0, swp1);
|
||||
return add1;
|
||||
# endif
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_cross(glm_vec4 v1, glm_vec4 v2)
|
||||
{
|
||||
glm_vec4 const swp0 = _mm_shuffle_ps(v1, v1, _MM_SHUFFLE(3, 0, 2, 1));
|
||||
glm_vec4 const swp1 = _mm_shuffle_ps(v1, v1, _MM_SHUFFLE(3, 1, 0, 2));
|
||||
glm_vec4 const swp2 = _mm_shuffle_ps(v2, v2, _MM_SHUFFLE(3, 0, 2, 1));
|
||||
glm_vec4 const swp3 = _mm_shuffle_ps(v2, v2, _MM_SHUFFLE(3, 1, 0, 2));
|
||||
glm_vec4 const mul0 = _mm_mul_ps(swp0, swp3);
|
||||
glm_vec4 const mul1 = _mm_mul_ps(swp1, swp2);
|
||||
glm_vec4 const sub0 = _mm_sub_ps(mul0, mul1);
|
||||
return sub0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_normalize(glm_vec4 v)
|
||||
{
|
||||
glm_vec4 const dot0 = glm_vec4_dot(v, v);
|
||||
glm_vec4 const isr0 = _mm_rsqrt_ps(dot0);
|
||||
glm_vec4 const mul0 = _mm_mul_ps(v, isr0);
|
||||
return mul0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_faceforward(glm_vec4 N, glm_vec4 I, glm_vec4 Nref)
|
||||
{
|
||||
glm_vec4 const dot0 = glm_vec4_dot(Nref, I);
|
||||
glm_vec4 const sgn0 = glm_vec4_sign(dot0);
|
||||
glm_vec4 const mul0 = _mm_mul_ps(sgn0, _mm_set1_ps(-1.0f));
|
||||
glm_vec4 const mul1 = _mm_mul_ps(N, mul0);
|
||||
return mul1;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_reflect(glm_vec4 I, glm_vec4 N)
|
||||
{
|
||||
glm_vec4 const dot0 = glm_vec4_dot(N, I);
|
||||
glm_vec4 const mul0 = _mm_mul_ps(N, dot0);
|
||||
glm_vec4 const mul1 = _mm_mul_ps(mul0, _mm_set1_ps(2.0f));
|
||||
glm_vec4 const sub0 = _mm_sub_ps(I, mul1);
|
||||
return sub0;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER __m128 glm_vec4_refract(glm_vec4 I, glm_vec4 N, glm_vec4 eta)
|
||||
{
|
||||
glm_vec4 const dot0 = glm_vec4_dot(N, I);
|
||||
glm_vec4 const mul0 = _mm_mul_ps(eta, eta);
|
||||
glm_vec4 const mul1 = _mm_mul_ps(dot0, dot0);
|
||||
glm_vec4 const sub0 = _mm_sub_ps(_mm_set1_ps(1.0f), mul0);
|
||||
glm_vec4 const sub1 = _mm_sub_ps(_mm_set1_ps(1.0f), mul1);
|
||||
glm_vec4 const mul2 = _mm_mul_ps(sub0, sub1);
|
||||
|
||||
if(_mm_movemask_ps(_mm_cmplt_ss(mul2, _mm_set1_ps(0.0f))) == 0)
|
||||
return _mm_set1_ps(0.0f);
|
||||
|
||||
glm_vec4 const sqt0 = _mm_sqrt_ps(mul2);
|
||||
glm_vec4 const mad0 = glm_vec4_fma(eta, dot0, sqt0);
|
||||
glm_vec4 const mul4 = _mm_mul_ps(mad0, N);
|
||||
glm_vec4 const mul5 = _mm_mul_ps(eta, I);
|
||||
glm_vec4 const sub2 = _mm_sub_ps(mul5, mul4);
|
||||
|
||||
return sub2;
|
||||
}
|
||||
|
||||
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
115
lib/glm/simd/integer.h
Normal file
115
lib/glm/simd/integer.h
Normal file
@@ -0,0 +1,115 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/integer.h
|
||||
|
||||
#pragma once
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_uvec4 glm_i128_interleave(glm_uvec4 x)
|
||||
{
|
||||
glm_uvec4 const Mask4 = _mm_set1_epi32(0x0000FFFF);
|
||||
glm_uvec4 const Mask3 = _mm_set1_epi32(0x00FF00FF);
|
||||
glm_uvec4 const Mask2 = _mm_set1_epi32(0x0F0F0F0F);
|
||||
glm_uvec4 const Mask1 = _mm_set1_epi32(0x33333333);
|
||||
glm_uvec4 const Mask0 = _mm_set1_epi32(0x55555555);
|
||||
|
||||
glm_uvec4 Reg1;
|
||||
glm_uvec4 Reg2;
|
||||
|
||||
// REG1 = x;
|
||||
// REG2 = y;
|
||||
//Reg1 = _mm_unpacklo_epi64(x, y);
|
||||
Reg1 = x;
|
||||
|
||||
//REG1 = ((REG1 << 16) | REG1) & glm::uint64(0x0000FFFF0000FFFF);
|
||||
//REG2 = ((REG2 << 16) | REG2) & glm::uint64(0x0000FFFF0000FFFF);
|
||||
Reg2 = _mm_slli_si128(Reg1, 2);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask4);
|
||||
|
||||
//REG1 = ((REG1 << 8) | REG1) & glm::uint64(0x00FF00FF00FF00FF);
|
||||
//REG2 = ((REG2 << 8) | REG2) & glm::uint64(0x00FF00FF00FF00FF);
|
||||
Reg2 = _mm_slli_si128(Reg1, 1);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask3);
|
||||
|
||||
//REG1 = ((REG1 << 4) | REG1) & glm::uint64(0x0F0F0F0F0F0F0F0F);
|
||||
//REG2 = ((REG2 << 4) | REG2) & glm::uint64(0x0F0F0F0F0F0F0F0F);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 4);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask2);
|
||||
|
||||
//REG1 = ((REG1 << 2) | REG1) & glm::uint64(0x3333333333333333);
|
||||
//REG2 = ((REG2 << 2) | REG2) & glm::uint64(0x3333333333333333);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 2);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask1);
|
||||
|
||||
//REG1 = ((REG1 << 1) | REG1) & glm::uint64(0x5555555555555555);
|
||||
//REG2 = ((REG2 << 1) | REG2) & glm::uint64(0x5555555555555555);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 1);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask0);
|
||||
|
||||
//return REG1 | (REG2 << 1);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 1);
|
||||
Reg2 = _mm_srli_si128(Reg2, 8);
|
||||
Reg1 = _mm_or_si128(Reg1, Reg2);
|
||||
|
||||
return Reg1;
|
||||
}
|
||||
|
||||
GLM_FUNC_QUALIFIER glm_uvec4 glm_i128_interleave2(glm_uvec4 x, glm_uvec4 y)
|
||||
{
|
||||
glm_uvec4 const Mask4 = _mm_set1_epi32(0x0000FFFF);
|
||||
glm_uvec4 const Mask3 = _mm_set1_epi32(0x00FF00FF);
|
||||
glm_uvec4 const Mask2 = _mm_set1_epi32(0x0F0F0F0F);
|
||||
glm_uvec4 const Mask1 = _mm_set1_epi32(0x33333333);
|
||||
glm_uvec4 const Mask0 = _mm_set1_epi32(0x55555555);
|
||||
|
||||
glm_uvec4 Reg1;
|
||||
glm_uvec4 Reg2;
|
||||
|
||||
// REG1 = x;
|
||||
// REG2 = y;
|
||||
Reg1 = _mm_unpacklo_epi64(x, y);
|
||||
|
||||
//REG1 = ((REG1 << 16) | REG1) & glm::uint64(0x0000FFFF0000FFFF);
|
||||
//REG2 = ((REG2 << 16) | REG2) & glm::uint64(0x0000FFFF0000FFFF);
|
||||
Reg2 = _mm_slli_si128(Reg1, 2);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask4);
|
||||
|
||||
//REG1 = ((REG1 << 8) | REG1) & glm::uint64(0x00FF00FF00FF00FF);
|
||||
//REG2 = ((REG2 << 8) | REG2) & glm::uint64(0x00FF00FF00FF00FF);
|
||||
Reg2 = _mm_slli_si128(Reg1, 1);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask3);
|
||||
|
||||
//REG1 = ((REG1 << 4) | REG1) & glm::uint64(0x0F0F0F0F0F0F0F0F);
|
||||
//REG2 = ((REG2 << 4) | REG2) & glm::uint64(0x0F0F0F0F0F0F0F0F);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 4);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask2);
|
||||
|
||||
//REG1 = ((REG1 << 2) | REG1) & glm::uint64(0x3333333333333333);
|
||||
//REG2 = ((REG2 << 2) | REG2) & glm::uint64(0x3333333333333333);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 2);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask1);
|
||||
|
||||
//REG1 = ((REG1 << 1) | REG1) & glm::uint64(0x5555555555555555);
|
||||
//REG2 = ((REG2 << 1) | REG2) & glm::uint64(0x5555555555555555);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 1);
|
||||
Reg1 = _mm_or_si128(Reg2, Reg1);
|
||||
Reg1 = _mm_and_si128(Reg1, Mask0);
|
||||
|
||||
//return REG1 | (REG2 << 1);
|
||||
Reg2 = _mm_slli_epi32(Reg1, 1);
|
||||
Reg2 = _mm_srli_si128(Reg2, 8);
|
||||
Reg1 = _mm_or_si128(Reg1, Reg2);
|
||||
|
||||
return Reg1;
|
||||
}
|
||||
|
||||
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
1028
lib/glm/simd/matrix.h
Normal file
1028
lib/glm/simd/matrix.h
Normal file
File diff suppressed because it is too large
Load Diff
8
lib/glm/simd/packing.h
Normal file
8
lib/glm/simd/packing.h
Normal file
@@ -0,0 +1,8 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/packing.h
|
||||
|
||||
#pragma once
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
452
lib/glm/simd/platform.h
Normal file
452
lib/glm/simd/platform.h
Normal file
@@ -0,0 +1,452 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/platform.h
|
||||
|
||||
#pragma once
|
||||
|
||||
///////////////////////////////////////////////////////////////////////////////////
|
||||
// Platform
|
||||
|
||||
#define GLM_PLATFORM_UNKNOWN 0x00000000
|
||||
#define GLM_PLATFORM_WINDOWS 0x00010000
|
||||
#define GLM_PLATFORM_LINUX 0x00020000
|
||||
#define GLM_PLATFORM_APPLE 0x00040000
|
||||
//#define GLM_PLATFORM_IOS 0x00080000
|
||||
#define GLM_PLATFORM_ANDROID 0x00100000
|
||||
#define GLM_PLATFORM_CHROME_NACL 0x00200000
|
||||
#define GLM_PLATFORM_UNIX 0x00400000
|
||||
#define GLM_PLATFORM_QNXNTO 0x00800000
|
||||
#define GLM_PLATFORM_WINCE 0x01000000
|
||||
#define GLM_PLATFORM_CYGWIN 0x02000000
|
||||
|
||||
#ifdef GLM_FORCE_PLATFORM_UNKNOWN
|
||||
# define GLM_PLATFORM GLM_PLATFORM_UNKNOWN
|
||||
#elif defined(__CYGWIN__)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_CYGWIN
|
||||
#elif defined(__QNXNTO__)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_QNXNTO
|
||||
#elif defined(__APPLE__)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_APPLE
|
||||
#elif defined(WINCE)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_WINCE
|
||||
#elif defined(_WIN32)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_WINDOWS
|
||||
#elif defined(__native_client__)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_CHROME_NACL
|
||||
#elif defined(__ANDROID__)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_ANDROID
|
||||
#elif defined(__linux)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_LINUX
|
||||
#elif defined(__unix)
|
||||
# define GLM_PLATFORM GLM_PLATFORM_UNIX
|
||||
#else
|
||||
# define GLM_PLATFORM GLM_PLATFORM_UNKNOWN
|
||||
#endif//
|
||||
|
||||
// Report platform detection
|
||||
#if GLM_MESSAGES == GLM_MESSAGES_ENABLED && !defined(GLM_MESSAGE_PLATFORM_DISPLAYED)
|
||||
# define GLM_MESSAGE_PLATFORM_DISPLAYED
|
||||
# if(GLM_PLATFORM & GLM_PLATFORM_QNXNTO)
|
||||
# pragma message("GLM: QNX platform detected")
|
||||
//# elif(GLM_PLATFORM & GLM_PLATFORM_IOS)
|
||||
//# pragma message("GLM: iOS platform detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_APPLE)
|
||||
# pragma message("GLM: Apple platform detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_WINCE)
|
||||
# pragma message("GLM: WinCE platform detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_WINDOWS)
|
||||
# pragma message("GLM: Windows platform detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_CHROME_NACL)
|
||||
# pragma message("GLM: Native Client detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_ANDROID)
|
||||
# pragma message("GLM: Android platform detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_LINUX)
|
||||
# pragma message("GLM: Linux platform detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_UNIX)
|
||||
# pragma message("GLM: UNIX platform detected")
|
||||
# elif(GLM_PLATFORM & GLM_PLATFORM_UNKNOWN)
|
||||
# pragma message("GLM: platform unknown")
|
||||
# else
|
||||
# pragma message("GLM: platform not detected")
|
||||
# endif
|
||||
#endif//GLM_MESSAGES
|
||||
|
||||
///////////////////////////////////////////////////////////////////////////////////
|
||||
// Compiler
|
||||
|
||||
#define GLM_COMPILER_UNKNOWN 0x00000000
|
||||
|
||||
// Intel
|
||||
#define GLM_COMPILER_INTEL 0x00100000
|
||||
#define GLM_COMPILER_INTEL12 0x00100010
|
||||
#define GLM_COMPILER_INTEL12_1 0x00100020
|
||||
#define GLM_COMPILER_INTEL13 0x00100030
|
||||
#define GLM_COMPILER_INTEL14 0x00100040
|
||||
#define GLM_COMPILER_INTEL15 0x00100050
|
||||
#define GLM_COMPILER_INTEL16 0x00100060
|
||||
|
||||
// Visual C++ defines
|
||||
#define GLM_COMPILER_VC 0x01000000
|
||||
#define GLM_COMPILER_VC10 0x01000090
|
||||
#define GLM_COMPILER_VC11 0x010000A0
|
||||
#define GLM_COMPILER_VC12 0x010000B0
|
||||
#define GLM_COMPILER_VC14 0x010000C0
|
||||
#define GLM_COMPILER_VC15 0x010000D0
|
||||
|
||||
// GCC defines
|
||||
#define GLM_COMPILER_GCC 0x02000000
|
||||
#define GLM_COMPILER_GCC44 0x020000B0
|
||||
#define GLM_COMPILER_GCC45 0x020000C0
|
||||
#define GLM_COMPILER_GCC46 0x020000D0
|
||||
#define GLM_COMPILER_GCC47 0x020000E0
|
||||
#define GLM_COMPILER_GCC48 0x020000F0
|
||||
#define GLM_COMPILER_GCC49 0x02000100
|
||||
#define GLM_COMPILER_GCC50 0x02000200
|
||||
#define GLM_COMPILER_GCC51 0x02000300
|
||||
#define GLM_COMPILER_GCC52 0x02000400
|
||||
#define GLM_COMPILER_GCC53 0x02000500
|
||||
#define GLM_COMPILER_GCC54 0x02000600
|
||||
#define GLM_COMPILER_GCC60 0x02000700
|
||||
#define GLM_COMPILER_GCC61 0x02000800
|
||||
#define GLM_COMPILER_GCC62 0x02000900
|
||||
#define GLM_COMPILER_GCC70 0x02000A00
|
||||
#define GLM_COMPILER_GCC71 0x02000B00
|
||||
#define GLM_COMPILER_GCC72 0x02000C00
|
||||
#define GLM_COMPILER_GCC80 0x02000D00
|
||||
|
||||
// CUDA
|
||||
#define GLM_COMPILER_CUDA 0x10000000
|
||||
#define GLM_COMPILER_CUDA40 0x10000040
|
||||
#define GLM_COMPILER_CUDA41 0x10000050
|
||||
#define GLM_COMPILER_CUDA42 0x10000060
|
||||
#define GLM_COMPILER_CUDA50 0x10000070
|
||||
#define GLM_COMPILER_CUDA60 0x10000080
|
||||
#define GLM_COMPILER_CUDA65 0x10000090
|
||||
#define GLM_COMPILER_CUDA70 0x100000A0
|
||||
#define GLM_COMPILER_CUDA75 0x100000B0
|
||||
#define GLM_COMPILER_CUDA80 0x100000C0
|
||||
|
||||
// Clang
|
||||
#define GLM_COMPILER_CLANG 0x20000000
|
||||
#define GLM_COMPILER_CLANG32 0x20000030
|
||||
#define GLM_COMPILER_CLANG33 0x20000040
|
||||
#define GLM_COMPILER_CLANG34 0x20000050
|
||||
#define GLM_COMPILER_CLANG35 0x20000060
|
||||
#define GLM_COMPILER_CLANG36 0x20000070
|
||||
#define GLM_COMPILER_CLANG37 0x20000080
|
||||
#define GLM_COMPILER_CLANG38 0x20000090
|
||||
#define GLM_COMPILER_CLANG39 0x200000A0
|
||||
#define GLM_COMPILER_CLANG40 0x200000B0
|
||||
#define GLM_COMPILER_CLANG41 0x200000C0
|
||||
#define GLM_COMPILER_CLANG42 0x200000D0
|
||||
|
||||
// Build model
|
||||
#define GLM_MODEL_32 0x00000010
|
||||
#define GLM_MODEL_64 0x00000020
|
||||
|
||||
// Force generic C++ compiler
|
||||
#ifdef GLM_FORCE_COMPILER_UNKNOWN
|
||||
# define GLM_COMPILER GLM_COMPILER_UNKNOWN
|
||||
|
||||
#elif defined(__INTEL_COMPILER)
|
||||
# if __INTEL_COMPILER == 1200
|
||||
# define GLM_COMPILER GLM_COMPILER_INTEL12
|
||||
# elif __INTEL_COMPILER == 1210
|
||||
# define GLM_COMPILER GLM_COMPILER_INTEL12_1
|
||||
# elif __INTEL_COMPILER == 1300
|
||||
# define GLM_COMPILER GLM_COMPILER_INTEL13
|
||||
# elif __INTEL_COMPILER == 1400
|
||||
# define GLM_COMPILER GLM_COMPILER_INTEL14
|
||||
# elif __INTEL_COMPILER == 1500
|
||||
# define GLM_COMPILER GLM_COMPILER_INTEL15
|
||||
# elif __INTEL_COMPILER >= 1600
|
||||
# define GLM_COMPILER GLM_COMPILER_INTEL16
|
||||
# else
|
||||
# define GLM_COMPILER GLM_COMPILER_INTEL
|
||||
# endif
|
||||
|
||||
// CUDA
|
||||
#elif defined(__CUDACC__)
|
||||
# if !defined(CUDA_VERSION) && !defined(GLM_FORCE_CUDA)
|
||||
# include <cuda.h> // make sure version is defined since nvcc does not define it itself!
|
||||
# endif
|
||||
# if CUDA_VERSION < 3000
|
||||
# error "GLM requires CUDA 3.0 or higher"
|
||||
# else
|
||||
# define GLM_COMPILER GLM_COMPILER_CUDA
|
||||
# endif
|
||||
|
||||
// Clang
|
||||
#elif defined(__clang__)
|
||||
# if defined(__apple_build_version__)
|
||||
# if __clang_major__ == 5 && __clang_minor__ == 0
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG33
|
||||
# elif __clang_major__ == 5 && __clang_minor__ == 1
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG34
|
||||
# elif __clang_major__ == 6 && __clang_minor__ == 0
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG35
|
||||
# elif __clang_major__ == 6 && __clang_minor__ >= 1
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG36
|
||||
# elif __clang_major__ >= 7
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG37
|
||||
# else
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG
|
||||
# endif
|
||||
# else
|
||||
# if __clang_major__ == 3 && __clang_minor__ == 0
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG30
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 1
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG31
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 2
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG32
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 3
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG33
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 4
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG34
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 5
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG35
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 6
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG36
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 7
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG37
|
||||
# elif __clang_major__ == 3 && __clang_minor__ == 8
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG38
|
||||
# elif __clang_major__ == 3 && __clang_minor__ >= 9
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG39
|
||||
# elif __clang_major__ == 4 && __clang_minor__ == 0
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG40
|
||||
# elif __clang_major__ == 4 && __clang_minor__ == 1
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG41
|
||||
# elif __clang_major__ == 4 && __clang_minor__ >= 2
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG42
|
||||
# elif __clang_major__ >= 4
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG42
|
||||
# else
|
||||
# define GLM_COMPILER GLM_COMPILER_CLANG
|
||||
# endif
|
||||
# endif
|
||||
|
||||
// Visual C++
|
||||
#elif defined(_MSC_VER)
|
||||
# if _MSC_VER < 1600
|
||||
# error "GLM requires Visual C++ 10 - 2010 or higher"
|
||||
# elif _MSC_VER == 1600
|
||||
# define GLM_COMPILER GLM_COMPILER_VC11
|
||||
# elif _MSC_VER == 1700
|
||||
# define GLM_COMPILER GLM_COMPILER_VC11
|
||||
# elif _MSC_VER == 1800
|
||||
# define GLM_COMPILER GLM_COMPILER_VC12
|
||||
# elif _MSC_VER == 1900
|
||||
# define GLM_COMPILER GLM_COMPILER_VC14
|
||||
# elif _MSC_VER >= 1910
|
||||
# define GLM_COMPILER GLM_COMPILER_VC15
|
||||
# else//_MSC_VER
|
||||
# define GLM_COMPILER GLM_COMPILER_VC
|
||||
# endif//_MSC_VER
|
||||
|
||||
// G++
|
||||
#elif defined(__GNUC__) || defined(__MINGW32__)
|
||||
# if (__GNUC__ == 4) && (__GNUC_MINOR__ == 2)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC42)
|
||||
# elif (__GNUC__ == 4) && (__GNUC_MINOR__ == 3)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC43)
|
||||
# elif (__GNUC__ == 4) && (__GNUC_MINOR__ == 4)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC44)
|
||||
# elif (__GNUC__ == 4) && (__GNUC_MINOR__ == 5)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC45)
|
||||
# elif (__GNUC__ == 4) && (__GNUC_MINOR__ == 6)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC46)
|
||||
# elif (__GNUC__ == 4) && (__GNUC_MINOR__ == 7)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC47)
|
||||
# elif (__GNUC__ == 4) && (__GNUC_MINOR__ == 8)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC48)
|
||||
# elif (__GNUC__ == 4) && (__GNUC_MINOR__ >= 9)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC49)
|
||||
# elif (__GNUC__ == 5) && (__GNUC_MINOR__ == 0)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC50)
|
||||
# elif (__GNUC__ == 5) && (__GNUC_MINOR__ == 1)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC51)
|
||||
# elif (__GNUC__ == 5) && (__GNUC_MINOR__ == 2)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC52)
|
||||
# elif (__GNUC__ == 5) && (__GNUC_MINOR__ == 3)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC53)
|
||||
# elif (__GNUC__ == 5) && (__GNUC_MINOR__ >= 4)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC54)
|
||||
# elif (__GNUC__ == 6) && (__GNUC_MINOR__ == 0)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC60)
|
||||
# elif (__GNUC__ == 6) && (__GNUC_MINOR__ == 1)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC61)
|
||||
# elif (__GNUC__ == 6) && (__GNUC_MINOR__ >= 2)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC62)
|
||||
# elif (__GNUC__ == 7) && (__GNUC_MINOR__ == 0)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC70)
|
||||
# elif (__GNUC__ == 7) && (__GNUC_MINOR__ == 1)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC71)
|
||||
# elif (__GNUC__ == 7) && (__GNUC_MINOR__ == 2)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC72)
|
||||
# elif (__GNUC__ >= 8)
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC80)
|
||||
# else
|
||||
# define GLM_COMPILER (GLM_COMPILER_GCC)
|
||||
# endif
|
||||
|
||||
#else
|
||||
# define GLM_COMPILER GLM_COMPILER_UNKNOWN
|
||||
#endif
|
||||
|
||||
#ifndef GLM_COMPILER
|
||||
# error "GLM_COMPILER undefined, your compiler may not be supported by GLM. Add #define GLM_COMPILER 0 to ignore this message."
|
||||
#endif//GLM_COMPILER
|
||||
|
||||
///////////////////////////////////////////////////////////////////////////////////
|
||||
// Instruction sets
|
||||
|
||||
// User defines: GLM_FORCE_PURE GLM_FORCE_SSE2 GLM_FORCE_SSE3 GLM_FORCE_AVX GLM_FORCE_AVX2 GLM_FORCE_AVX2
|
||||
|
||||
#define GLM_ARCH_X86_BIT 0x00000001
|
||||
#define GLM_ARCH_SSE2_BIT 0x00000002
|
||||
#define GLM_ARCH_SSE3_BIT 0x00000004
|
||||
#define GLM_ARCH_SSSE3_BIT 0x00000008
|
||||
#define GLM_ARCH_SSE41_BIT 0x00000010
|
||||
#define GLM_ARCH_SSE42_BIT 0x00000020
|
||||
#define GLM_ARCH_AVX_BIT 0x00000040
|
||||
#define GLM_ARCH_AVX2_BIT 0x00000080
|
||||
#define GLM_ARCH_AVX512_BIT 0x00000100 // Skylake subset
|
||||
#define GLM_ARCH_ARM_BIT 0x00000100
|
||||
#define GLM_ARCH_NEON_BIT 0x00000200
|
||||
#define GLM_ARCH_MIPS_BIT 0x00010000
|
||||
#define GLM_ARCH_PPC_BIT 0x01000000
|
||||
|
||||
#define GLM_ARCH_PURE (0x00000000)
|
||||
#define GLM_ARCH_X86 (GLM_ARCH_X86_BIT)
|
||||
#define GLM_ARCH_SSE2 (GLM_ARCH_SSE2_BIT | GLM_ARCH_X86)
|
||||
#define GLM_ARCH_SSE3 (GLM_ARCH_SSE3_BIT | GLM_ARCH_SSE2)
|
||||
#define GLM_ARCH_SSSE3 (GLM_ARCH_SSSE3_BIT | GLM_ARCH_SSE3)
|
||||
#define GLM_ARCH_SSE41 (GLM_ARCH_SSE41_BIT | GLM_ARCH_SSSE3)
|
||||
#define GLM_ARCH_SSE42 (GLM_ARCH_SSE42_BIT | GLM_ARCH_SSE41)
|
||||
#define GLM_ARCH_AVX (GLM_ARCH_AVX_BIT | GLM_ARCH_SSE42)
|
||||
#define GLM_ARCH_AVX2 (GLM_ARCH_AVX2_BIT | GLM_ARCH_AVX)
|
||||
#define GLM_ARCH_AVX512 (GLM_ARCH_AVX512_BIT | GLM_ARCH_AVX2) // Skylake subset
|
||||
#define GLM_ARCH_ARM (GLM_ARCH_ARM_BIT)
|
||||
#define GLM_ARCH_NEON (GLM_ARCH_NEON_BIT | GLM_ARCH_ARM)
|
||||
#define GLM_ARCH_MIPS (GLM_ARCH_MIPS_BIT)
|
||||
#define GLM_ARCH_PPC (GLM_ARCH_PPC_BIT)
|
||||
|
||||
#if defined(GLM_FORCE_PURE)
|
||||
# define GLM_ARCH GLM_ARCH_PURE
|
||||
#elif defined(GLM_FORCE_MIPS)
|
||||
# define GLM_ARCH (GLM_ARCH_MIPS)
|
||||
#elif defined(GLM_FORCE_PPC)
|
||||
# define GLM_ARCH (GLM_ARCH_PPC)
|
||||
#elif defined(GLM_FORCE_NEON)
|
||||
# define GLM_ARCH (GLM_ARCH_NEON)
|
||||
#elif defined(GLM_FORCE_AVX512)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX512)
|
||||
#elif defined(GLM_FORCE_AVX2)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX2)
|
||||
#elif defined(GLM_FORCE_AVX)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX)
|
||||
#elif defined(GLM_FORCE_SSE42)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE42)
|
||||
#elif defined(GLM_FORCE_SSE41)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE41)
|
||||
#elif defined(GLM_FORCE_SSSE3)
|
||||
# define GLM_ARCH (GLM_ARCH_SSSE3)
|
||||
#elif defined(GLM_FORCE_SSE3)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE3)
|
||||
#elif defined(GLM_FORCE_SSE2)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE2)
|
||||
#elif (GLM_COMPILER & (GLM_COMPILER_CLANG | GLM_COMPILER_GCC)) || ((GLM_COMPILER & GLM_COMPILER_INTEL) && (GLM_PLATFORM & GLM_PLATFORM_LINUX))
|
||||
// This is Skylake set of instruction set
|
||||
# if defined(__AVX512BW__) && defined(__AVX512F__) && defined(__AVX512CD__) && defined(__AVX512VL__) && defined(__AVX512DQ__)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX512)
|
||||
# elif defined(__AVX2__)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX2)
|
||||
# elif defined(__AVX__)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX)
|
||||
# elif defined(__SSE4_2__)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE42)
|
||||
# elif defined(__SSE4_1__)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE41)
|
||||
# elif defined(__SSSE3__)
|
||||
# define GLM_ARCH (GLM_ARCH_SSSE3)
|
||||
# elif defined(__SSE3__)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE3)
|
||||
# elif defined(__SSE2__)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE2)
|
||||
# elif defined(__i386__) || defined(__x86_64__)
|
||||
# define GLM_ARCH (GLM_ARCH_X86)
|
||||
# elif defined(__ARM_NEON)
|
||||
# define GLM_ARCH (GLM_ARCH_ARM | GLM_ARCH_NEON)
|
||||
# elif defined(__arm__ )
|
||||
# define GLM_ARCH (GLM_ARCH_ARM)
|
||||
# elif defined(__mips__ )
|
||||
# define GLM_ARCH (GLM_ARCH_MIPS)
|
||||
# elif defined(__powerpc__ )
|
||||
# define GLM_ARCH (GLM_ARCH_PPC)
|
||||
# else
|
||||
# define GLM_ARCH (GLM_ARCH_PURE)
|
||||
# endif
|
||||
#elif (GLM_COMPILER & GLM_COMPILER_VC) || ((GLM_COMPILER & GLM_COMPILER_INTEL) && (GLM_PLATFORM & GLM_PLATFORM_WINDOWS))
|
||||
# if defined(_M_ARM)
|
||||
# define GLM_ARCH (GLM_ARCH_ARM)
|
||||
# elif defined(__AVX2__)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX2)
|
||||
# elif defined(__AVX__)
|
||||
# define GLM_ARCH (GLM_ARCH_AVX)
|
||||
# elif defined(_M_X64)
|
||||
# define GLM_ARCH (GLM_ARCH_SSE2)
|
||||
# elif defined(_M_IX86_FP)
|
||||
# if _M_IX86_FP >= 2
|
||||
# define GLM_ARCH (GLM_ARCH_SSE2)
|
||||
# else
|
||||
# define GLM_ARCH (GLM_ARCH_PURE)
|
||||
# endif
|
||||
# elif defined(_M_PPC)
|
||||
# define GLM_ARCH (GLM_ARCH_PPC)
|
||||
# else
|
||||
# define GLM_ARCH (GLM_ARCH_PURE)
|
||||
# endif
|
||||
#else
|
||||
# define GLM_ARCH GLM_ARCH_PURE
|
||||
#endif
|
||||
|
||||
// With MinGW-W64, including intrinsic headers before intrin.h will produce some errors. The problem is
|
||||
// that windows.h (and maybe other headers) will silently include intrin.h, which of course causes problems.
|
||||
// To fix, we just explicitly include intrin.h here.
|
||||
#if defined(__MINGW64__) && (GLM_ARCH != GLM_ARCH_PURE)
|
||||
# include <intrin.h>
|
||||
#endif
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_AVX2_BIT
|
||||
# include <immintrin.h>
|
||||
#elif GLM_ARCH & GLM_ARCH_AVX_BIT
|
||||
# include <immintrin.h>
|
||||
#elif GLM_ARCH & GLM_ARCH_SSE42_BIT
|
||||
# if GLM_COMPILER & GLM_COMPILER_CLANG
|
||||
# include <popcntintrin.h>
|
||||
# endif
|
||||
# include <nmmintrin.h>
|
||||
#elif GLM_ARCH & GLM_ARCH_SSE41_BIT
|
||||
# include <smmintrin.h>
|
||||
#elif GLM_ARCH & GLM_ARCH_SSSE3_BIT
|
||||
# include <tmmintrin.h>
|
||||
#elif GLM_ARCH & GLM_ARCH_SSE3_BIT
|
||||
# include <pmmintrin.h>
|
||||
#elif GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
# include <emmintrin.h>
|
||||
#endif//GLM_ARCH
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
typedef __m128 glm_vec4;
|
||||
typedef __m128i glm_ivec4;
|
||||
typedef __m128i glm_uvec4;
|
||||
#endif
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_AVX_BIT
|
||||
typedef __m256d glm_dvec4;
|
||||
#endif
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_AVX2_BIT
|
||||
typedef __m256i glm_i64vec4;
|
||||
typedef __m256i glm_u64vec4;
|
||||
#endif
|
||||
9
lib/glm/simd/trigonometric.h
Normal file
9
lib/glm/simd/trigonometric.h
Normal file
@@ -0,0 +1,9 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/trigonometric.h
|
||||
|
||||
#pragma once
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
8
lib/glm/simd/vector_relational.h
Normal file
8
lib/glm/simd/vector_relational.h
Normal file
@@ -0,0 +1,8 @@
|
||||
/// @ref simd
|
||||
/// @file glm/simd/vector_relational.h
|
||||
|
||||
#pragma once
|
||||
|
||||
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
|
||||
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
|
||||
Reference in New Issue
Block a user