baram2584/FESADev
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Integrate Refactor solver sources and dual MSVC CMake builds

Browse Source
This commit is contained in:
KOKO\Mimi
2026-10-18 02:10:51 +09:00
parent 7ee32e5837
commit 12d51480de
218 changed files with 17842 additions and 9689 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/fesa/elements/hex8_pair_simd_dispatch.cpp
+13 -13
View File
@@ -1,6 +1,6 @@
#include "hex8_pair_simd.h"
#include "hex8_pair_arithmetic.h"
#include "../math/high_low_arithmetic.h"
#if defined(FESA_HEX8_PAIR_AVX2) && defined(_MSC_VER) && defined(_M_X64)
#include <float.h>
@@ -16,11 +16,11 @@ namespace fesa::hex8_internal {
#endif
FESA_HEX8_BASELINE_NOINLINE
void ScalarPacketAdd(const PairPacket& left, const PairPacket& right,
bool pair_shortcuts, PairPacket* output) {
PairPacket* output) {
for (std::size_t lane = 0; lane < 4; ++lane) {
const math_internal::HighLow a{left.values[lane].high, left.values[lane].low};
const math_internal::HighLow b{right.values[lane].high, right.values[lane].low};
const auto value = pair_shortcuts ? AddPair(a, b) : math_internal::Add(a, b);
const auto value = math_internal::Add(a, b);
output->values[lane] = {value.high, value.low};
}
}
@@ -28,7 +28,7 @@ FESA_HEX8_BASELINE_NOINLINE
void ScalarPacketMultiply(const PairPacket& left, const PairPacket& right,
PairPacket* output) {
for (std::size_t lane = 0; lane < 4; ++lane) {
const auto value = MultiplyPair(
const auto value = math_internal::Multiply(
{left.values[lane].high, left.values[lane].low},
{right.values[lane].high, right.values[lane].low});
output->values[lane] = {value.high, value.low};
@@ -51,14 +51,14 @@ void EvaluateGeometryPacketScalar(const GeometryPacketInput& input,
for (std::size_t lane = 0; lane < 4; ++lane) {
math_internal::HighLow dot;
for (std::size_t i = 0; i < 3; ++i)
dot = AddPair(dot, MultiplyPair(value(input.gradient_a[i]),
dot = math_internal::Add(dot, math_internal::Multiply(value(input.gradient_a[i]),
value(input.gradient_b[i].values[lane])));
math_internal::HighLow geo;
for (std::size_t j = 0; j < 3; ++j)
for (std::size_t l = 0; l < 3; ++l)
geo = AddPair(geo, MultiplyPair(value(input.gradient_stress[j][l]),
geo = math_internal::Add(geo, math_internal::Multiply(value(input.gradient_stress[j][l]),
value(input.gradient_b[l].values[lane])));
const auto weighted = MultiplyPair(geo, value(input.weight));
const auto weighted = math_internal::Multiply(geo, value(input.weight));
output->gradient_dot.values[lane] = {dot.high, dot.low};
output->weighted_geometric.values[lane] = {weighted.high, weighted.low};
}
@@ -84,15 +84,15 @@ void EvaluateMaterialPacketScalar(const MaterialPacketInput& input,
return math_internal::HighLow{pair.high, pair.low};
};
for (std::size_t lane = 0; lane < 4; ++lane) {
const auto first = MultiplyPair(value(input.deviatoric_left.values[lane]),
const auto first = math_internal::Multiply(value(input.deviatoric_left.values[lane]),
value(input.deviatoric_right.values[lane]));
const auto second = MultiplyPair(value(input.shear_left.values[lane]),
const auto second = math_internal::Multiply(value(input.shear_left.values[lane]),
value(input.shear_right.values[lane]));
const auto third = MultiplyPair(value(input.tensor_left.values[lane]),
const auto third = math_internal::Multiply(value(input.tensor_left.values[lane]),
value(input.gradient_dot.values[lane]));
const auto material = AddPair(AddPair(first, second), third);
const auto accumulated = AddPair(value(input.accumulated.values[lane]),
MultiplyPair(material, value(input.weight)));
const auto material = math_internal::Add(math_internal::Add(first, second), third);
const auto accumulated = math_internal::Add(value(input.accumulated.values[lane]),
math_internal::Multiply(material, value(input.weight)));
output->values[lane] = {accumulated.high, accumulated.low};
}
}