Method init_mappings

smallthinker/src/llama-model-loader.cpp:825–856

}

void llama_model_loader::init_mappings(bool prefetch, llama_mlocks * mlock_mmaps) {
    if (use_mmap) {
        mappings.reserve(files.size());
        mmaps_used.reserve(files.size());
        for (const auto & file : files) {
            bool is_numa = false;

            auto * dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU);
            if (dev) {
                auto * reg = ggml_backend_dev_backend_reg(dev);
                auto * is_numa_fn = (decltype(ggml_is_numa) *) ggml_backend_reg_get_proc_address(reg, "ggml_backend_cpu_is_numa");
                if (is_numa_fn) {
                    is_numa = is_numa_fn();
                }
            }

            std::unique_ptr<llama_mmap> mapping = std::make_unique<llama_mmap>(file.get(), prefetch ? -1 : 0, is_numa);
            mmaps_used.emplace_back(mapping->size(), 0);
            if (mlock_mmaps) {
                std::unique_ptr<llama_mlock> mlock_mmap(new llama_mlock());
                mlock_mmap->init(mapping->addr());
                mlock_mmaps->emplace_back(std::move(mlock_mmap));
            }
            mappings.emplace_back(std::move(mapping));
        }
    }

    // compute the total size of all tensors for progress reporting
    for (const auto & it : weights_map) {
        size_data += ggml_nbytes(it.second.tensor);
    }
}

void llama_model_loader::get_mapping_range(size_t * first, size_t * last, void ** addr, int idx, ggml_context * ctx) const {
    GGML_ASSERT(!mappings.empty());