FCloud

CMakeLists.txt ```cmake cmake_minimum_required(VERSION 3.25) set(CMAKE_CUDA_ARCHITECTURES 80 86) set(CMAKE_CUDA_STANDARD 17) SET(CMAKE_CUDA_COMPILER /usr/local/cuda/bin/nvcc) project(demo_test CUDA CXX) find_package(CUDA REQUIRED) find_package(CUDAToolkit REQUIRED) include_directories(/usr/local/cuda/include) include_directories(/usr/include) aux_source_directory(src SOURCES) include(cmake/CPM.cmake) CPMAddPackage( NAME bght URL "https://github.com/owensgroup/BGHT/archive/refs/heads/main.zip" OPTIONS "build_tests...

Yes There are lots of threads try to insert the hash map. I reference the code to edit my kernel: ```cuda template __global__ void createLookupHashKernel(HashMap lookupTable, const int *coordinates, int...

It works for insert process, so if I also have a kernel function that reads hash map in multi threads, i need to use same operations?

I use this to find values by tables we created before: ` createRulesTableByHashKernel(HashMap lookupTable, const int *coordinates, int N, int *rulesTable, int max_x, int max_y, int max_z, int kernel_size) {...

I am confused with InputIt first and InputIt last in examples, I only need to implement one search at one time. I noticed: ` template __global__ void find_kernel(InputIt first, InputIt...