README.md

Standard Library for CUDA Programming

This library provides data structures to ease programming in CUDA (version 12 or higher). For a tutorial and further information, please read this manual.

Example

Quick example on how to transfer a std::vector on CPU to a battery::vector on GPU (notice you don't need to do any manual memory allocation or deallocation):

#include <vector>
#include "battery/vector.hpp"
#include "battery/unique_ptr.hpp"
#include "battery/allocator.hpp"

using mvector = battery::vector<int, battery::managed_allocator>;

__global__ void kernel(mvector* v_ptr) {
  mvector& v = *v_ptr;
  // ... Compute on `v` in parallel.
}

int main(int argc, char** argv) {
  std::vector<int> v(10000, 42);
  // Transfer from CPU vector to GPU vector.
  auto gpu_v = battery::make_unique<mvector, battery::managed_allocator>(v);
  kernel<<<256, 256>>>(gpu_v.get());
  CUDAEX(cudaDeviceSynchronize());
  // Transfering the new data to the initial vector.
  for(int i = 0; i < v.size(); ++i) {
    v[i] = (*gpu_v)[i];
  }
  return 0;
}

Common Questions

• How to transfer data from the CPU to the GPU?
• How to create a CMake project for CUDA project?
• How to allocate a vector shared by all threads of a block inside a kernel?
• How to allocate a vector shared by all blocks inside a kernel?
• CUDA runtime error an illegal memory access was encountered
• How to allocate a vector in shared memory?

Quick Reference

• Namespace: battery::*.
• The documentation is not exhaustive (which is why we provide a link to the standard C++ STL documentation), but we document most of the main differences and the features without a standard counterpart.
• The table below is a quick reference to the most useful features, but it is not exhaustive.
• The structures provided here are not thread-safe, this responsibility is delegated to the user of this library.

Category	Main features
Allocator	standard_allocator	global_allocator	managed_allocator	pool_allocator
Pointers	shared_ptr (std)	make_shared (std)	allocate_shared (std)
	unique_ptr (std)	make_unique (std)	make_unique_block	make_unique_grid
Containers	vector (std)	string (std)	dynamic_bitset
	tuple	variant (std)	bitset (std)
Utility	CUDA	INLINE	CUDAE	CUDAEX
	limits	ru_cast	rd_cast
	popcount (std)	countl_zero (std)	countl_one (std)	countr_zero (std)
	countr_one (std)	signum	ipow
	add_up	add_down	sub_up	sub_down
	mul_up	mul_down	div_up	div_down
Memory	local_memory	read_only_memory	atomic_memory
	atomic_scoped_memory	atomic_memory_block	atomic_memory_grid