Class Kernel

gpu.hpp:455–463

* @brief Represents handles + metadata for a reusable kernel on the GPU. * The struct members can be divided into "consumed upon dispatch" * (commandBuffer) and reusable ahead-of-time setup (all other members). */

 * (commandBuffer) and reusable ahead-of-time setup (all other members).
 */
struct Kernel {
  std::unique_ptr<WGPUBuffer[]> buffers; // non-owning
  std::unique_ptr<size_t[]> bufferSizes;
  size_t numBindings;
  Shape totalWorkgroups;
  WGPUBindGroup bindGroup;             // persists between submission
  WGPUComputePipeline computePipeline; // persists between submission
  WGPUCommandBuffer commandBuffer;     // destroyed upon submission
};


/**