randomkang commented on code in PR #3144: URL: https://github.com/apache/brpc/pull/3144#discussion_r3879252777
########## src/butil/gpu/gpu_block_pool.cpp: ########## @@ -0,0 +1,509 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, +// software distributed under the License is distributed on an +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +// KIND, either express or implied. See the License for the +// specific language governing permissions and limitations +// under the License. + +#if BRPC_WITH_GDR + +#include <iostream> +#include <chrono> +#include <gflags/gflags.h> +#include <sys/stat.h> +#include "butil/fast_rand.h" +#include "gpu_block_pool.h" +namespace butil { +namespace gdr { +DEFINE_int32(gdr_block_size_kb, 512, "gdr block size in KB"); +DEFINE_int32(max_gdr_regions, 32, "max num of gdr regions"); + +#define CHECK_CUDA(call) \ +do { \ + auto _sts = (call); \ + if (_sts != cudaSuccess) { \ + LOG(FATAL) << " cuda error:" \ + << (cudaGetErrorString(_sts)) << std::string(" at ") \ + << __FILE__ << ": " << __LINE__; \ + } \ +} while (0); + +size_t GetGdrBlockSize() { + return FLAGS_gdr_block_size_kb * 1024; +} + +bool verify_same_context() { + static std::once_flag init_flag; + static int original_device = -1; + + int current_device; + CHECK_CUDA(cudaGetDevice(¤t_device)); + + std::call_once(init_flag, [&]() { + original_device = current_device; + }); + + return (current_device == original_device); +} + +void* get_gpu_mem(int gpu_id, int64_t gpu_mem_size) { + CHECK_CUDA(cudaSetDevice(gpu_id)); + void *d_data; + + LOG(INFO) << "try to alloc " << gpu_mem_size << " bytes from gpu " << gpu_id; + + CHECK_CUDA(cudaMalloc(&d_data, gpu_mem_size)); + cudaDeviceSynchronize(); + return (void *)d_data; +} + +void* get_cpu_mem(int gpu_id, int64_t cpu_mem_size) { + CHECK_CUDA(cudaSetDevice(gpu_id)); + + LOG(INFO) << "try to alloc " << cpu_mem_size << " bytes from gpu " << gpu_id << "on host"; + + void* mem = NULL; + + CHECK_CUDA(cudaMallocHost(&mem, cpu_mem_size)); + + cudaDeviceSynchronize(); + + return mem; +} + + +std::atomic<BlockPoolAllocators*> BlockPoolAllocators::instance_{nullptr}; + +BlockPoolAllocators* BlockPoolAllocators::singleton() { Review Comment: 我看了下，可以用brpc自带的Singleton，已改 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
