Skip to content
Tao Luo edited this page Dec 9, 2019 · 1 revision

6-13 Need discussion this week

  • MKL dynamic link @tangjian下周会开始看
  • parallelDo 的性能差于v2, 需要磨平吗
  • 需要push NLP上线了吗

6-13 周报

@tangjian

性能

  • 现在v2的性能可以跟以前磨平了,但是fluid没有跟之前磨平。

NLP

MKLDNN

5-31 Need discussion this week

  • ResNet50在6148上的训练和预测性能,与Intel Team对齐(QA)@chengsi 先拿一个短期的机器
  • inference示例加入repo并能在ci上验证 https://github.com/PaddlePaddle/Paddle/issues/10990#issuecomment-393034634
  • 将contrib/inference_api打包并部署
  • 6148 机器
  • QA 人力不够
  • Train CPU with multi-thread @luotao 参考parallel do
  • TODO Inference with multi-thread @tangjian done

5-31 会议纪要

  • 6148机器 ready
  • MKLDNN 7.5 的 milestone 可以设成与 V2 对齐
  • CPU 训练多线程尝试下 ParallelDo
  • Inference 7.5 目标
    • 高层 API 及文档
      • 原生实现
      • Anakin 集成
      • 子图初步集成 TensorRT
    • CPU 两个模型达到上线标准
    • 子图初步完成
      • 目标为 MLP ,争取 resnet
    • 图像 4个 inference demo

5-31 周报

Inference engine && op 整体优化

整体进展

  • 高层API 60%
    • DONE 确定高层接口
    • DONE 原生实现
    • DOING 阿拉金
  • 子图 30%
    • DOING framework
    • DOING TRT 支持
  • MKLDNN 30%
  • CPU核心模型优化 70%
    • DOING OCR CPU
    • DOING 情感分类 CPU
  • 文档 && CI 40%
    • DONE 旧接口文档初步
    • TODO 新接口文档

本周进展

MKLDNN

高层API

@panxin

@chunwei

CPU核心模型优化

  • OCR: 使用mklml动态库会导致其他任务coredump @luotao @visualization-team
    • [已验证生效] 因为在使用MKLDNN和MKLML的时候会需要iomp.so,所以线上也必须要链接iomp。但是线上有很多服务已经使用了gomp,我们需要推荐用户直接使用iomp,链接方式是:target_link_libraries(${TARGET_NAME} "-L${MKLML_LIB_DIR} -liomp5 -Wl,--as-needed") @tangjian @yangjingyuan-OCR
    • [正在进行中] 依然存在mklml动态库与其他任务的mkl静态库冲突的问题。尝试使用MKL大包中的-lmkl_core -lmkl_intel_lp64 -lmkl_sequential进行编译,尝试v1版本的情况。@luotao @yangjingyuan-OCR @lixuan-OCR
    • [正在沟通中] 向Intel MKL组的 ying.hu@intel.com 寻求单线程和多线程版的静态MKLML库
  • 情感分类@tensor
    • 正在出性能报告

文档 && CI版本

下周计划

  • 高层API 正式发布(接口稳定,文档基本ready) @chunwei @panxin
  • 子图人工配置整体跑通 @chunwei
  • CPU核心模型出性能报告以及解决线上bug(联合 Intel团队)@luotao @tensor

release-note

  • 下周提供高层API的release note @chunwei

5-23 待讨论

5-23 周报

MKLDNN 进展

  • 确认和波兰团队的双周会议,暂定下周三下午一点为第一次会议。
  • MKLDNN的时间规划,波兰团队邮件中说这周末给出(他们经理Marcin在外面参加会议)。
    • we are merging internally the latest code, supporting MKL-DNN data layouts, elementwise_add and SUM to check, how are we doing with Resnet training/inference and to identify next bottlenecks.
    • CRNN-CTC model shall also benefit from those upgrades. From OPs point of view, we will look at top_k layer and optimize algorithm by parallelizing it (as it is a sorting problem, MKL-DNN isn’t a good place to implement it)
  • Merged add mkldnn to paddle lib @qiaolongfei
  • Merged Reuse of pooling mkldnn primitives @intel team
  • Merged Update activations for MKL-DNN @intel team
  • Merged enable MKLDNN inference test @tangjian

CPU核心模型优化

子图Engine进展

高层接口封装

版本发布 & 文档

CI 编译

存在问题:

  • 包有多余路径:paddle/build/fluid_install_dir
  • GIT COMMIT ID在CI中没有打出来
  • 其他版本的部署

文档

Bug fix

5-23 会议纪要

  • MKLDNN 如何与Intel团队沟通产出/承诺有疑惑,需要@wangyi确定
  • save_inference_model 需要接受targets 留空,默认输出所有逻辑(除了backward)

5-16 会议纪要

计划讨论

  • Refactor inference documentation and deploy on the official website
  • tracking the status of different branches
  • tracking the status of deployment seven models
  • the plan of MKLDNN, more clear milestones
  • refine the GitHub/projects
  • determine the date of the internal weekly meeting
  • have a GitHub wiki for tracking the common questions and bugs
  • Release the Inference Lib with a document containing version details such as the following information so that users can reproduce it
    • commit id
    • the compilation commands and flags
    • a performance report (QA provides)
  • Higher API to hide the underlying details (including concepts, third-party symbols)

文档方面

建立文档机制,初期可以让 shanyi 帮把现有文档迁移到 paddle/doc,同时开始跑通部署到 official site 的流程

周会机制

每周三晚上 6:30~7:30,Inference内部讨论,增加远程协同的信息带宽

周会目的是,回顾过去一周的工作,讨论共享信息,以及确定下周大概的工作目标和方向

其中,会有如下流程

面向 7 个上线模型的优化

  • 明确上线的目标性能
  • 确定模型优先级
  • 每个模型,有沟通好完整的测试性能的方法,对应的测试数据等,自己可以复现业务线的指标
  • 分析瓶颈,逐步优化

Inference 框架

  • 优化框架,Kernel 的重复创建可以cache下,这点在 cudnn和 mkldnn上比较明显,具体收益不太确定
  • 高层API,与图像沟通,他们会按照我们现在的高层API,先将他们的接口实现贡献过来,上层用很薄一层封装;未来我们的版本发布需要对性能问题负责

CPU 方向

  • MKLDNN需要用单线程测试结果
  • MKLML 对比 Openblas 有明显性能增益,可以考虑先替代 openblas 拿到 CPU的第一步收益
  • 对于MKLDNN还未支持,但又是特定模型瓶颈的 OP,可以自己使用类似MKLDNN的方式优化(这点可能sys也是这么做的)

GPU 方向

  • 尽早完成子图测试验证
Clone this wiki locally