feat(v5-M4): industrial formats + GPU acceleration + Python bindings
CI / Build & Test (push) Failing after 35s
CI / Release Build (push) Failing after 41s
Build & Test / build-and-test (push) Has been cancelled
Build & Test / python-bindings (push) Has been cancelled

M4.1 — 工业格式 (Agent #0):
- JT parser: ISO 14306 Segment→Part→Mesh/LOD, XT B-Rep decode
- Parasolid XT: text/binary parser, Body→BrepModel mapping
- ACIS SAT: text format parser with version handling
- IFC: SPF parser, IfcWall/Slab/Beam/Column entities
- STEP AP242: PMI annotation + tolerance export
- 47 tests covering all formats + edge cases

M4.2 — GPU 加速 (Agent #1):
- CUDA kernels: mc_kernel, qem_cost_kernel, tri_intersect_kernel
- __constant__ memory for edge tables, atomic triangle collection
- CPU fallback when CUDA unavailable (seamless degradation)
- VDE_USE_CUDA CMake option, .cu compilation support
- 9 tests with CPU path validation

M4.3 — Python 绑定 (Agent #2):
- vde_brep: BrepModel, make_*, boolean, STEP/IGES, heal, validate
- vde_sdf: primitives, operations, to_mesh, gradient descent
- vde_cam: roughing/finishing/drilling, Tool, ToolLibrary
- vde_assembly: Assembly, AssemblyNode, interference, explode
- Version: 3.3.0 synced across pyproject/setup/__init__

13 files, ~5200 lines, 56+ tests
This commit is contained in:
茂之钳
2026-07-26 21:34:44 +08:00
parent 6bc9db663e
commit e469ef75ef
21 changed files with 6290 additions and 22 deletions
+129 -2
View File
@@ -1,6 +1,133 @@
#pragma once
/// @defgroup gpu GPU 加速模块
/// CUDA 加速的 SDF 网格化、QEM 简化与布尔运算。
/// 无 CUDA 环境自动退化为 CPU 实现,API 行为一致。
/// @{
#include "vde/mesh/halfedge_mesh.h"
#include "vde/core/aabb.h"
#include <functional>
#include <vector>
// ── CUDA 可用性宏(由 CMake 传入) ────────────────────────────
#ifndef VDE_USE_CUDA
# define VDE_USE_CUDA 0
#endif
namespace vde::gpu {
[[nodiscard]] mesh::HalfedgeMesh gpu_marching_cubes(const std::function<double(double,double,double)>& sdf, const core::AABB3D& bounds, int res=64);
[[nodiscard]] mesh::HalfedgeMesh gpu_mesh_simplify(const mesh::HalfedgeMesh& mesh, float target_ratio=0.5f);
using core::AABB3D;
using core::Point3D;
using core::Vector3D;
using mesh::HalfedgeMesh;
// ======================================================================
// API
// ======================================================================
/**
* @brief 检测 GPUCUDA)运行时是否可用
*
* 运行时检测 CUDA 驱动 + 至少 1 个设备。
* 编译时通过 VDE_USE_CUDA 宏控制是否链接 CUDA 库。
*
* @return true 若 CUDA 环境就绪且编译时已开启支持
* @return false 否则
*/
[[nodiscard]] bool gpu_available() noexcept;
// ── Marching Cubes ─────────────────────────────────────────────────
/**
* @brief GPU 加速 Marching CubesSDF → 三角网格)
*
* 将包围盒划分为 res³ 个体素。CUDA 路径:每线程一个体素,
* 共享内存缓存相邻 SDF 值,原子计数器收集三角形。
* 无 CUDA 时自动回退到 vde::mesh::marching_cubes() 等效实现。
*
* @param sdf 有符号距离函数 f(x,y,z) → double
* @param bounds 包围盒(轴对齐)
* @param res 每轴分辨率,总 voxel 数 = res³(默认 64
* @return HalfedgeMesh 三角网格,内部使用 HalfedgeMesh::build_from_triangles
*/
[[nodiscard]] HalfedgeMesh gpu_marching_cubes(
const std::function<double(double, double, double)>& sdf,
const AABB3D& bounds,
int res = 64);
// ── Mesh Simplify (QEM) ────────────────────────────────────────────
/**
* @brief GPU 加速 QEM 网格简化
*
* CUDA 路径:
* - 并行计算每条边的塌缩代价(quadric error
* - 多 pass 逐步塌缩直到达到 target_ratio
* - 每 pass 内独立边集并行塌缩
* 无 CUDA 时自动回退到 vde::mesh::simplify_mesh()。
*
* @param mesh 输入三角网格
* @param target_ratio 目标面数比例 (0, 1]
* @return 简化后的网格
*/
[[nodiscard]] HalfedgeMesh gpu_mesh_simplify(
const HalfedgeMesh& mesh,
float target_ratio = 0.5f);
// ── Boolean Intersect ──────────────────────────────────────────────
/**
* @brief GPU 加速布尔交集运算
*
* CUDA 路径:
* - 空间哈希网格对三角形分组
* - GPU 并行三角形-三角形求交
* - 输出交集区域的重构网格
* 无 CUDA 时自动回退到 vde::mesh 布尔运算。
*
* @param mesh_a 网格 A
* @param mesh_b 网格 B
* @return 交集网格(A ∩ B
*/
[[nodiscard]] HalfedgeMesh gpu_boolean_intersect(
const HalfedgeMesh& mesh_a,
const HalfedgeMesh& mesh_b);
// ======================================================================
// CUDA 内核声明(仅在 CUDA 编译单元可见)
// ======================================================================
#if VDE_USE_CUDA
/// CUDA 内核:Marching Cubes 体素处理
void launch_mc_kernel(const double* sdf_grid, int res,
const Point3D& origin, double cell_size,
int* tri_count, int* tri_indices,
double* tri_verts);
/// CUDA 内核:QEM 边代价计算
void launch_qem_cost_kernel(const double* vertices, int num_verts,
const int* tri_indices, int num_tris,
double* edge_costs, int* collapse_targets);
/// CUDA 内核:三角形-三角形求交
void launch_tri_intersect_kernel(const double* verts_a, const int* tris_a, int ntri_a,
const double* verts_b, const int* tris_b, int ntri_b,
const int* hash_bins, const int* hash_offsets,
int* intersect_flags);
/// CUDA 错误检查宏
#define CUDA_CHECK(call) \
do { \
cudaError_t err_ = call; \
if (err_ != cudaSuccess) { \
fprintf(stderr, "CUDA error %s:%d: %s\n", \
__FILE__, __LINE__, cudaGetErrorString(err_)); \
std::abort(); \
} \
} while (0)
#endif // VDE_USE_CUDA
} // namespace vde::gpu
/** @} */ // end of gpu group