Chapter 13 • Blyx Unique Feature
GPU Compute Blocks
Inline gpu {} blocks lower directly to SPIR-V and NVPTX GPU assembly without external C++ CUDA host wrappers.
gpu {
let tid = thread_id();
data[tid] = data[tid] * 2.0;
};Chapter 13 • Blyx Unique Feature
Inline gpu {} blocks lower directly to SPIR-V and NVPTX GPU assembly without external C++ CUDA host wrappers.
gpu {
let tid = thread_id();
data[tid] = data[tid] * 2.0;
};