CUDA Backend¶
Note
The CUDA/GPU backend is controlled by the SPARQ_ENABLE_CUDA switch (OFF by default; only the CPU
path is compiled). Enabling it requires a local CUDA toolchain (CUDA 13 / CCCL 3 tested). GPU kernels
for the production CondRot primitives are not yet available; the relevant operators automatically
fall back to the CPU path on GPU states.
GPU Basic Components (SparQ/include/cuda/basic_components.cuh)¶
GPU Conditional Rotations (SparQ/include/cuda/condrot.cuh)¶
GPU Interference Basic Components (SparQ/include/cuda/quantum_interfere_basic.cuh)¶
GPU Sparse State Simulator Entry (SparQ/include/cuda/sparse_state_simulator.cuh)¶
中文版 ===
CUDA 后端¶
Note
CUDA/GPU 后端由 SPARQ_ENABLE_CUDA 开关控制（默认 OFF，仅编译 CPU 路径；
开启需本机 CUDA 工具链，CUDA 13 / CCCL 3 实测通过）。生产 CondRot 原语的
GPU 内核尚未提供，相关算子在 GPU 态自动回退 CPU 路径执行。