|
|
cb8d8eefe4
|
修改了lal_gran_hooke.cpp中resize函数的输入错误，避免幽灵粒子情况下的内存越界
|
2025-07-13 22:57:19 +08:00 |
|
|
|
82f0d869bc
|
lal_gran_hooke.cpp文件的resize函数的输入有bug，调整后可以正常运行幽灵粒子相关算例
|
2025-07-13 22:52:01 +08:00 |
|
|
|
2774fc7b10
|
只修改了一个resize的数据内存空间重新分配的地方
|
2025-07-13 14:34:03 +08:00 |
|
|
|
ec60fcf1fe
|
build(gpu): 添加CUDA调试支持并扩展编译选项
* 在GPU.cmake中添加了${CUDA_NVCC_FLAGS_CUSTOM}变量到CUDA编译命令
* 引入gpu-granular-debug.cmake构建预设配置
* 添加调试模式支持(-g -G -O0)和详细的编译器检测
* 为所有构建模式设置-allow-unsupported-compiler标志
|
2025-07-10 15:07:17 +08:00 |
|
|
|
9978b23eef
|
feat(granregion): 重构漏斗示例并添加两阶段仿真流程
* 修改 in.granregion.funnel:
- 将颗粒尺寸从 [0.25, 0.5] 调整为 [0.2, 0.3]
- 接触模型从 hertz/history 更改为 hooke
- 增加单次注入颗粒数量从 2000 到 8000
- 添加邻域列表设置 neigh_modify one 10000
* 新增两阶段仿真文件:
- in.granregion_init.funnel: 初始化阶段，颗粒注入并生成重启文件
- in.granregion_run.funnel: 运行阶段，从重启文件继续仿真
* 其他调整：
- 去除初始化阶段的颗粒注入步骤注释
- 添加边界区域重新定义逻辑
- 优化热力学数据输出设置
|
2025-07-10 09:21:16 +08:00 |
|
|
|
6dcfc82be5
|
feat(gpu/pair): 添加 Hooke 颗粒力模型的 GPU 加速支持
实现了颗粒间 Hooke 力模型的 GPU 加速计算，主要包含以下内容：
* 引入基础类 BaseSphere 处理球形颗粒的位置、半径和角速度数据
* 新增 GranHooke 类实现法向/切向弹性力与阻尼力的 GPU 计算逻辑
* 使用纹理内存优化颗粒属性访问（半径、质量、速度、角速度）
* 实现设备端邻居列表构建和并行力计算内核
* 支持与 LAMMPS 的 FixNeighHistory 集成处理接触历史数据
* 添加 GPU 与 CPU 数据转换接口，支持混合粒子分组和冻结状态处理
* 整合到 LAMMPS 的 GPU 模块，可通过 pair_style gran/hooke/gpu 调用
|
2025-07-07 14:37:56 +08:00 |
|
|
|
63fcdb6e52
|
Merge pull request #4634 from akohlmey/static-analysis-updates
Updates for LAMMPS from static code analysis
|
2025-06-29 22:25:15 -04:00 |
|
|
|
cd44539429
|
modernize by using anonymous namespace
|
2025-06-29 19:56:42 -04:00 |
|
|
|
47332f8e46
|
no need to make member const
|
2025-06-29 19:28:00 -04:00 |
|
|
|
fee96bc777
|
make double to bool/int/tagint conversions explicit
|
2025-06-29 19:10:29 -04:00 |
|
|
|
b279ba9a40
|
follow one definition rule by moving functions into class definition
|
2025-06-29 18:47:40 -04:00 |
|
|
|
0c68c4d493
|
Merge pull request #4636 from wapisani/fix_bond_create_inter_intra
Add support for inter-/intra-molecular bonding to fix bond/create
|
2025-06-28 09:43:54 -04:00 |
|
|
|
3e12c3f773
|
flag move constructor and swap operator as noexcept for performance
|
2025-06-28 05:21:25 -04:00 |
|
|
|
7bc86899b3
|
type fixes and error message updates
|
2025-06-28 05:20:49 -04:00 |
|
|
|
ac7013432b
|
don't use double as boolean, use qualified auto
|
2025-06-28 04:48:47 -04:00 |
|
|
|
039fb94faa
|
improve errors and apply clang-format
|
2025-06-28 04:45:02 -04:00 |
|
|
|
8c64b2090f
|
process tilt parameters correctly and consistent with FixNH class
|
2025-06-28 04:38:10 -04:00 |
|
|
|
77417e2223
|
use correct type
|
2025-06-28 04:37:24 -04:00 |
|
|
|
2e4f29c900
|
use correct type and converions function
|
2025-06-28 04:24:46 -04:00 |
|
|
|
8dae514e03
|
use correct types
|
2025-06-28 04:06:36 -04:00 |
|
|
|
22a42446d3
|
tap_flag is integer and thus we should use utils::inumeric()
|
2025-06-28 03:58:27 -04:00 |
|
|
|
2968bfa8d9
|
nsvector_total should be an int
|
2025-06-28 03:53:34 -04:00 |
|
|
|
669f60b692
|
use qualified auto and avoid redundant types
|
2025-06-28 03:51:09 -04:00 |
|
|
|
1bdf30103f
|
used correct type parser and make conversions explicit
|
2025-06-28 03:50:51 -04:00 |
|
|
|
4992caed95
|
more modernization of DIFFRACTION package
|
2025-06-28 03:37:42 -04:00 |
|
|
|
f425f2ab26
|
modernize
|
2025-06-28 03:30:10 -04:00 |
|
|
|
309e374867
|
use utils::inumeric() for parsing integers, using qualified auto
|
2025-06-28 03:19:17 -04:00 |
|
|
|
d5a344a5c8
|
modernize
|
2025-06-28 03:18:52 -04:00 |
|
|
|
481749dc35
|
modernize argument processing and error messages
|
2025-06-28 03:07:17 -04:00 |
|
|
|
14b1241db2
|
use utils::inumeric() for reading integer arguments
|
2025-06-28 03:05:03 -04:00 |
|
|
|
4d26bd5fd5
|
ndouble is an int
|
2025-06-28 03:04:44 -04:00 |
|
|
|
5fe4c9149f
|
use qualified auto
|
2025-06-28 03:04:33 -04:00 |
|
|
|
a24aeb3fc6
|
icut, jcut should be doubles
|
2025-06-28 03:04:23 -04:00 |
|
|
|
0cd4ea4d04
|
don't use doubles as boolean
|
2025-06-28 02:34:58 -04:00 |
|
|
|
8726deb9b6
|
newton_bond is int not double
|
2025-06-28 02:34:39 -04:00 |
|
|
|
46561f57bd
|
use utils::inumeric() to parse integers
|
2025-06-28 02:34:22 -04:00 |
|
|
|
aacc30c4d7
|
no benefit from declaring _tid const
|
2025-06-28 02:33:55 -04:00 |
|
|
|
2833b7adbd
|
make cast to double explicit
|
2025-06-28 02:33:38 -04:00 |
|
|
|
11c4a8b49c
|
use math.h function without std:: prefix
|
2025-06-28 02:22:25 -04:00 |
|
|
|
c28869124b
|
don't truncate cuberoot to single precision
|
2025-06-28 02:22:09 -04:00 |
|
|
|
bd98354691
|
make cast to double precision explicit
|
2025-06-28 02:17:25 -04:00 |
|
|
|
109aec74ea
|
remove one more implicit truncation of precision. update unit test again.
|
2025-06-28 02:06:49 -04:00 |
|
|
|
588ed11e85
|
avoid comparison with different signedness
|
2025-06-28 02:01:48 -04:00 |
|
|
|
be5602c44f
|
apply clang-format
|
2025-06-28 02:01:32 -04:00 |
|
|
|
e78d85725b
|
convert to internal linkage with anonymous namespace
|
2025-06-28 02:01:23 -04:00 |
|
|
|
1b36451f6a
|
modernize. turn static inline into just inline
|
2025-06-27 23:58:11 -04:00 |
|
|
|
d168de00db
|
silence compiler warnings
|
2025-06-27 23:57:37 -04:00 |
|
|
|
0c6ce8f23c
|
fix typo
|
2025-06-27 23:27:44 -04:00 |
|
|
|
3ffe858a8b
|
remove dead code and protect from neighbor list with special neighbors
this addresses most of issue #4639
|
2025-06-27 23:25:15 -04:00 |
|
|
|
c279d194dc
|
be paranoid about invalid molecule templates
|
2025-06-27 22:47:50 -04:00 |
|
