DEME 2.0.1__tar.gz → 2.0.3__tar.gz

{deme-2.0.1 → deme-2.0.3}/CMakeLists.txt

@@ -35,6 +35,8 @@ set(CMAKE_USE_WIN32_THREADS_INIT 0)
 set(CMAKE_USE_PTHREADS_INIT 1)
 set(THREADS_PREFER_PTHREAD_FLAG ON)
 
+# Mark compile time (not runtime), affects how some kernel files include
+add_compile_definitions(DEME_BEING_CMAKE_COMPILED)
 
 # ---------------------------------------------------------------------------- #
 # Additional Packages

{deme-2.0.1/deme.egg-info → deme-2.0.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: deme
-Version: 2.0.1
+Version: 2.0.3
 Summary: PyBind Wrapper Library for DEM-Engine
 Home-page: UNKNOWN
 Author: Ruochun Zhang

{deme-2.0.1 → deme-2.0.3/deme.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: deme
-Version: 2.0.1
+Version: 2.0.3
 Summary: PyBind Wrapper Library for DEM-Engine
 Home-page: UNKNOWN
 Author: Ruochun Zhang

{deme-2.0.1 → deme-2.0.3}/setup.py

@@ -6,6 +6,7 @@ from pathlib import Path
 
 from setuptools import Extension, setup
 from pybind11.setup_helpers import Pybind11Extension, build_ext
+
 # from setuptools.command.build_ext import build_ext
 
 # Convert distutils Windows platform specifiers to CMake -A arguments
@@ -23,12 +24,21 @@ PLAT_TO_CMAKE = {
 
 class CMakeExtension(Pybind11Extension):
     def __init__(self, name: str, sourcedir: str = "") -> None:
-
         # Obtaining conda_prefix environment variable
         conda_prefix = os.environ.get("CONDA_PREFIX")
 
-        super().__init__(name, sources=[], extra_link_args=['--no-as-needed', '-Wl,-rpath,' + conda_prefix + '/lib/python3.10/site-packages/lib'], runtime_library_dirs=[conda_prefix +
-                                                                                                                                                                         '/lib/python3.10/site-packages/lib'], libraries=["libDEMERuntimeDataHelper"], include_dirs=["./src/DEM/", "./src/DEM/VariableTypes"], extra_objects=["./build"])
+        py_version = f"{sys.version_info.major}.{sys.version_info.minor}"
+        rpath = f"{conda_prefix}/lib/python{py_version}/site-packages/lib"
+
+        super().__init__(
+            name,
+            sources=[],
+            extra_link_args=["--no-as-needed", f"-Wl,-rpath,{rpath}"],
+            runtime_library_dirs=[rpath],
+            libraries=["libDEMERuntimeDataHelper"],
+            include_dirs=["./src/DEM/", "./src/DEM/VariableTypes"],
+            extra_objects=["./build"],
+        )
         self.sourcedir = os.fspath(Path(sourcedir).resolve())
 
 
@@ -37,14 +47,12 @@ class CMakeBuild(build_ext):
         # Must be in this form due to bug in .resolve() only fixed in Python 3.10+
         ext_fullpath = Path.cwd() / self.get_ext_fullpath(ext.name)
 
-        extdir = os.path.abspath(os.path.dirname(
-            self.get_ext_fullpath(ext.name)))
+        extdir = os.path.abspath(os.path.dirname(self.get_ext_fullpath(ext.name)))
 
         # Using this requires trailing slash for auto-detection & inclusion of
         # auxiliary "native" libs
 
-        debug = int(os.environ.get("DEBUG", 0)
-                    ) if self.debug is None else self.debug
+        debug = int(os.environ.get("DEBUG", 0)) if self.debug is None else self.debug
         cfg = "Debug" if debug else "Release"
 
         # CMake lets you override the generator - we need to check this.
@@ -67,12 +75,10 @@ class CMakeBuild(build_ext):
         # Adding CMake arguments set as environment variable
         # (needed e.g. to build for ARM OSx on conda-forge)
         if "CMAKE_ARGS" in os.environ:
-            cmake_args += [
-                item for item in os.environ["CMAKE_ARGS"].split(" ") if item]
+            cmake_args += [item for item in os.environ["CMAKE_ARGS"].split(" ") if item]
 
         # In this example, we pass in the version to C++. You might not need to.
-        cmake_args += [
-            f"-DEXAMPLE_VERSION_INFO={self.distribution.get_version()}"]
+        cmake_args += [f"-DEXAMPLE_VERSION_INFO={self.distribution.get_version()}"]
 
         if self.compiler.compiler_type != "msvc":
             # Using Ninja-build since it a) is available as a wheel and b)
@@ -94,8 +100,7 @@ class CMakeBuild(build_ext):
 
         else:
             # Single config generators are handled "normally"
-            single_config = any(
-                x in cmake_generator for x in {"NMake", "Ninja"})
+            single_config = any(x in cmake_generator for x in {"NMake", "Ninja"})
 
             # CMake allows an arch-in-generator style for backward compatibility
             contains_arch = any(x in cmake_generator for x in {"ARM", "Win64"})
@@ -117,8 +122,7 @@ class CMakeBuild(build_ext):
             # Cross-compile support for macOS - respect ARCHFLAGS if set
             archs = re.findall(r"-arch (\S+)", os.environ.get("ARCHFLAGS", ""))
             if archs:
-                cmake_args += [
-                    "-DCMAKE_OSX_ARCHITECTURES={}".format(";".join(archs))]
+                cmake_args += ["-DCMAKE_OSX_ARCHITECTURES={}".format(";".join(archs))]
 
         # Set CMAKE_BUILD_PARALLEL_LEVEL to control the parallel build level
         # across all generators.
@@ -148,7 +152,7 @@ class CMakeBuild(build_ext):
 # logic and declaration, and simpler if you include description/version in a file.
 setup(
     name="deme",
-    version="2.0.1",
+    version="2.0.3",
     author="Ruochun Zhang",
     author_email="ruochunz@gmail.com",
     description="PyBind Wrapper Library for DEM-Engine",
@@ -158,6 +162,6 @@ setup(
     zip_safe=False,
     include_package_data=True,
     sources=["./src/"],
-    install_requires=['numpy', 'pybind11'],
-    extras_require={"dev": ['pybind11'], "test": ["pytest>=6.0"]},
+    install_requires=["numpy", "pybind11"],
+    extras_require={"dev": ["pybind11"], "test": ["pytest>=6.0"]},
 )

{deme-2.0.1 → deme-2.0.3}/src/DEM/APIPrivate.cpp

@@ -302,19 +302,30 @@ void DEMSolver::jitifyKernels() {
     equipForceModel(m_subs);
     equipIntegrationScheme(m_subs);
     equipKernelIncludes(m_subs);
-    kT->jitifyKernels(m_subs);
-    dT->jitifyKernels(m_subs);
 
-    // Now, inspectors need to be jitified too... but the current design jitify inspector kernels at the first time they
-    // are used. for (auto& insp : m_inspectors) {
-    //     insp->Initialize(m_subs);
-    // }
-
-    // Solver system's own max vel inspector should be init-ed. Don't bother init-ing it while using, because it is
-    // called at high frequency, let's save an if check. Forced initialization (since doing it before system completes
-    // init).
-    m_approx_max_vel_func->Initialize(m_subs, true);
-    dT->approxMaxVelFunc = m_approx_max_vel_func;
+    // Jitify may require a defined device to derive the arch
+    std::thread kT_build([&]() {
+        DEME_GPU_CALL(cudaSetDevice(kT->streamInfo.device));
+        kT->jitifyKernels(m_subs);
+    });
+
+    std::thread dT_build([&]() {
+        DEME_GPU_CALL(cudaSetDevice(dT->streamInfo.device));
+        dT->jitifyKernels(m_subs);
+
+        // Now, inspectors need to be jitified too... but the current design jitify inspector kernels at the first time
+        // they are used. for (auto& insp : m_inspectors) {
+        //     insp->Initialize(m_subs);
+        // }
+
+        // Solver system's own max vel inspector should be init-ed. Don't bother init-ing it while using, because it is
+        // called at high frequency, let's save an if check. Forced initialization (since doing it before system
+        // completes init).
+        m_approx_max_vel_func->Initialize(m_subs, true);
+        dT->approxMaxVelFunc = m_approx_max_vel_func;
+    });
+    kT_build.join();
+    dT_build.join();
 }
 
 void DEMSolver::getContacts_impl(std::vector<bodyID_t>& idA,

{deme-2.0.1 → deme-2.0.3}/src/DEM/APIPublic.cpp

@@ -2191,12 +2191,16 @@ void DEMSolver::UpdateSimParams() {
     // Now that all params prepared, we need to migrate that imformation to the device
     migrateSimParamsToDevice();
 
-    // Jitify max vel finder, in case the policy there changed
-    m_approx_max_vel_func->Initialize(m_subs, true);
-    dT->approxMaxVelFunc = m_approx_max_vel_func;
-
-    // Updating sim environment is critical
-    dT->announceCritical();
+    // JItify may require a defined device to decide the arch
+    std::thread dT_build([&]() {
+        DEME_GPU_CALL(cudaSetDevice(dT->streamInfo.device));
+        // Jitify max vel finder, in case the policy there changed
+        m_approx_max_vel_func->Initialize(m_subs, true);
+        dT->approxMaxVelFunc = m_approx_max_vel_func;
+        // Updating sim environment is critical
+        dT->announceCritical();
+    });
+    dT_build.join();
 }
 
 void DEMSolver::UpdateStepSize(double ts) {
@@ -2310,11 +2314,9 @@ void DEMSolver::ChangeClumpSizes(const std::vector<bodyID_t>& IDs, const std::ve
 void DEMSolver::PurgeFamily(unsigned int family_num) {}
 
 void DEMSolver::DoDynamics(double thisCallDuration) {
-    // Is it needed here??
-    // dT->packDataPointers(kT->granData);
-
-    // TODO: Return if nSphere == 0
-    // TODO: Check if initialized
+    if (!sys_initialized) {
+        Initialize();
+    }
 
     // Tell dT how long this call is
     dT->setCycleDuration(thisCallDuration);

{deme-2.0.1 → deme-2.0.3}/src/algorithms/CMakeLists.txt

@@ -7,8 +7,6 @@ message(STATUS "==== Configuring GPU Algorithm Components ====")
 
 set (algorithms_message "[algorithms]")
 
-
-
 add_library(algorithms OBJECT)
 
 target_include_directories(

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_BallDrop.py

@@ -125,5 +125,5 @@ if __name__ == "__main__":
         t += frame_time
 
     DEMSim.ShowTimingStats()
-    DEMSim.ShowAnomalies()
+    DEMSim.ShowMemStats()
     print("BallDrop demo exiting...")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_BallDrop2D.py

@@ -318,6 +318,8 @@ if __name__ == "__main__":
 
         DEMSim.ShowTimingStats()
 
+    DEMSim.ShowMemStats()
+
     final_pos = proj_tracker.Pos()
     print(f"Ball density: {ball_density}")
     print(f"Ball rad: {R}")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_Centrifuge.py

@@ -181,6 +181,6 @@ if __name__ == "__main__":
     DEMSim.ClearThreadCollaborationStats()
 
     DEMSim.ShowTimingStats()
-    DEMSim.ShowAnomalies()
+    DEMSim.ShowMemStats()
 
     print("DEMdemo_Centrifuge exiting...")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_ConePenetration.py

@@ -267,4 +267,5 @@ if __name__ == "__main__":
         frame_count += 1
         t += frame_time
 
+    DEMSim.ShowMemStats()
     print(f"ConePenetration demo exiting...")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_Electrostatic.py

@@ -392,4 +392,5 @@ if __name__ == "__main__":
         rod_tracker.SetPos([0, 0, rod_length / 2. + current_height])
 
     DEMSim.ShowTimingStats()
+    DEMSim.ShowMemStats()
     print("Electrostatic demo exiting...")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_FlexibleMesh.py

@@ -272,4 +272,5 @@ if __name__ == "__main__":
         step_count += 1
 
     DEMSim.ShowTimingStats()
+    DEMSim.ShowMemStats()
     print("FlexibleMesh demo exiting...")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_GRCPrep_Part1.py

@@ -157,6 +157,7 @@ if __name__ == "__main__":
 
     DEMSim.ShowThreadCollaborationStats()
     DEMSim.ClearThreadCollaborationStats()
+    DEMSim.ShowMemStats()
 
     cp_filename = os.path.join(out_dir, f"GRC_3e5.csv")
     cnt_filename = os.path.join(out_dir, f"Contact_pairs_3e5.csv")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_GameOfLife.py

@@ -123,4 +123,6 @@ if __name__ == "__main__":
 
         DEMSim.DoDynamicsThenSync(1.)
 
+    DEMSim.ShowMemStats()
+
     print("DEMdemo_GameOfLife exiting...")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_Mixer.py

@@ -131,7 +131,7 @@ if __name__ == "__main__":
 
         mixer_moi = np.array(mixer_tracker.MOI())
         mixer_acc = np.array(mixer_tracker.ContactAngAccLocal())
-        mixer_torque = np.cross(mixer_acc, mixer_moi)
+        mixer_torque = mixer_acc * mixer_moi
         print(
             f"Contact torque on the mixer is {mixer_torque[0]}, {mixer_torque[1]}, {mixer_torque[2]}", flush=True)
 
@@ -144,4 +144,5 @@ if __name__ == "__main__":
     print(f"{elapsed_time} seconds (wall time) to finish this simulation")
 
     DEMSim.ShowTimingStats()
+    DEMSim.ShowMemStats()
     print("Mixer demo exiting...")

{deme-2.0.1 → deme-2.0.3}/src/demo/pyDEME_WheelDPSimplified.py

@@ -218,6 +218,6 @@ if __name__ == "__main__":
         t += step_size
 
     DEMSim.ShowTimingStats()
-    DEMSim.ShowAnomalies()
+    DEMSim.ShowMemStats()
 
     print("WheelDPSimpilified demo exiting...")

{deme-2.0.1 → deme-2.0.3}/src/kernel/DEMHelperKernels.cuh

@@ -4,7 +4,14 @@
 #define DEME_HELPER_KERNELS_CU
 
 #include <DEM/Defines.h>
-#include <kernel/CUDAMathHelpers.cuh>
+
+// CUDAMathHelpers.cuh is used in static compilation, and that needs to be found through src then kernel. But at
+// runtime, we ask jitify to find include files only at same level of all other fellow kernel files.
+#ifdef DEME_BEING_CMAKE_COMPILED
+    #include <kernel/CUDAMathHelpers.cuh>
+#else
+    #include <CUDAMathHelpers.cuh>
+#endif
 
 // inline __device__ voxelID_t position2VoxelID