【信息科学与工程学】【人工智能】百万上下文大语言模型算法02

// 预取 __builtin_prefetch(&data[i + prefetch_distance * 8], 0, 3); } // 处理当前缓存行 for (int j = 0; j < 8 && i + j < size; j++) { process_element(data[i + j]); } } break; } case AccessPattern::RANDOM: { // 随机访问，使用软件预取 int64_t* data = tensor.data_ptr<int64_t>(); int64_t size = tensor.numel(); // 生成访问模式 std::vector<int> indices(size); std::iota(indices.begin(), indices.end(), 0); std::random_shuffle(indices.begin(), indices.end()); int prefetch_ahead = 8; for (size_t i = 0;