1. 03 11月, 2021 2 次提交
  2. 18 10月, 2021 1 次提交
    • M
      feat(mgb/gopt): add profiler cache · 012de769
      Megvii Engine Team 提交于
      In order to improve performance of the profiling procedure. Make layout transform testcase stable. The profiling result in ci environment will be cached in files.
      
      GitOrigin-RevId: ba2743f35fcdbd554b7cd82e70f433ccdcc66fa4
      012de769
  3. 01 10月, 2021 1 次提交
  4. 19 9月, 2021 1 次提交
  5. 18 9月, 2021 1 次提交
  6. 18 8月, 2021 1 次提交
  7. 03 8月, 2021 2 次提交
    • M
      feat(bazel/windows/xp/sp2/inference): implement inference on windows xp · c68e6695
      Megvii Engine Team 提交于
      (os vesion >= sp2) build with bazel
      
      * bazel build support(define __DEPLOY_ON_XP_SP2__ when deploy on xp sp2):
      (dbg)./bazel build //brain/megbrain:load_and_run --cpu='x86_windows_xp'
      --compiler='clang_cl' -c dbg --copt "-D__DEPLOY_ON_XP_SP2__=1"
      
      (opt)./bazel build //brain/megbrain:load_and_run --cpu='x86_windows_xp'
      --compiler='clang_cl' -c opt --copt "-D__DEPLOY_ON_XP_SP2__=1"
      
      * internal behavior:
      will define MGB_HAVE_THREAD=0 when enable __DEPLOY_ON_XP_SP2__
      
      * refer to
      https://docs.microsoft.com/en-us/cpp/build/configuring-programs-for-windows-xp?view=msvc-160
      xp sp2(x86) do not support vc runtime fully, casused by KERNEL32.dll do not
      implement some base apis for c++ std function, for example,
      std::mutex/std::thread/std::condition_variable as a workround, we will
      disable some MegEngine features on xp sp2 env, for exampe, multi-thread etc!
      
      * about DNN_MUTEX/MGB_MUTEX, if your code will build in inference
      code (even CPU backends), please replace std::mutex to DNN_MUTEX/MGB_MUTEX,
      
      * about multi-thread, if you code need multi-thread support, please
      enable it when MGB_HAVE_THREAD=1
      
      * about test build env status
      1: Visual Studio 2019(MSVC version <= 14.26.28801)---- pass
      2: Visual Studio 2019(MSVC version > 14.26.28801) ---- failed
         caused by this 'new' version will put VCR depends on win7
         KERNEL32.DLL, this may be fixed at Visual Studio 2019 later version
         but we do not test at this MR merge point
      3: Visual Studio 2017   ---------- pass
      4: Visual Studio 2014   ---------- pass
      GitOrigin-RevId: 65ac48b95e99f2c510fe5db449cc8182d682e113
      c68e6695
    • M
      feat(cmake/windows/cuda): upgrade windows llvm to · 125406e3
      Megvii Engine Team 提交于
      lastest 12.0.1, for fit cutlass upgrade
      
      GitOrigin-RevId: 23b9d779dfbfc6dbbf31b63abdcc7a4aa6c654cd
      125406e3
  8. 19 7月, 2021 7 次提交
  9. 21 6月, 2021 2 次提交
  10. 03 6月, 2021 3 次提交
  11. 25 5月, 2021 2 次提交
  12. 14 4月, 2021 1 次提交
  13. 23 3月, 2021 2 次提交
  14. 22 3月, 2021 1 次提交
  15. 02 2月, 2021 2 次提交
  16. 23 1月, 2021 1 次提交
    • M
      fix(cmake/whl): reduce wheel size · ae8b38f6
      Megvii Engine Team 提交于
      as some reason: some version of trt/cudnn need
      -Wl,--whole-archive attribute to fix cuda issue,
      target A(CMAKE OBJECT,eg megbrain/megdnn): -Wl,--whole-archive depend on trt/cudnn
      target B(SHARED LIBS,eg libmegengine.so) --> A with `PUBLIC` depends
      target C(SHARED LIBS,eg _impertive.so) --> B with `PUBLIC` depends,
      then ld will force link fatbin section into C, which will be undesired
      
      what`s more, attribute PUBLIC/PRIVATE do not take effect
      to OBJECT library(megbrain/megdnn)
      
      what`s more, megengine/megengine_export have to PUBLIC for
      mgb/imperative target, as SRC/include depends
      
      so we pull cudalib depend from megbrain/megdnn to
      megengine/megengine_export on linux os, to finall
      target at windows os, for example lar on windows
      
      GitOrigin-RevId: b278a69e1cdcc17cfd22da80247af3fe30f6d4e8
      ae8b38f6
  17. 11 1月, 2021 5 次提交
  18. 11 12月, 2020 2 次提交
  19. 04 12月, 2020 1 次提交
  20. 05 11月, 2020 2 次提交