Commits · 293eef6df428c12e2d2b8eac5b5b4c84d3ebeb42 · OpenDAS / FastMoE

28 Jan, 2021 1 commit
- hack nccl of pytorch · 293eef6d
  Rick Ho authored Jan 28, 2021
  
  293eef6d
25 Jan, 2021 1 commit
- move original files to examples · 0c47afbb
  Rick Ho authored Jan 25, 2021
  
  0c47afbb
10 Jan, 2021 1 commit
- global scatter gather kernels and pytorch C function · ec322e4b
  Rick Ho authored Jan 10, 2021
  
  ec322e4b
09 Jan, 2021 4 commits
- stream manager across multiple devices · 60b93e39
  Rick Ho authored Jan 09, 2021
  
  60b93e39
- make backward pass test · 27d8beaa
  Rick Ho authored Jan 09, 2021
  
  27d8beaa
- split operators and make forward run · 49732231
  Rick Ho authored Jan 09, 2021
  
  49732231
- stream manager fixed · 2565f2fa
  Rick Ho authored Jan 09, 2021
  
  2565f2fa
04 Jan, 2021 1 commit
- stream manager object instead of pointer · ef83c893
  Jiezhong Qiu authored Jan 04, 2021
  
  ef83c893
03 Jan, 2021 3 commits
- fix memory leak · e52c0380
  Jiezhong Qiu authored Jan 03, 2021
  
  e52c0380
- update · 6212aa28
  Jiezhong Qiu authored Jan 03, 2021
  
  6212aa28
- make cudaStreamManager thread local · 27af1828
  Jiezhong Qiu authored Jan 03, 2021
  
  27af1828
02 Jan, 2021 2 commits
- CUDA error at moe_cuda_kernel.cu:86 code=13(CUBLAS_STATUS_EXECUTION_FAILED) · 85306aa7
  Jiezhong Qiu authored Jan 03, 2021
  
  85306aa7
- set device according to input · 046455a8
  Jiezhong Qiu authored Jan 03, 2021
  
  046455a8
30 Dec, 2020 4 commits
- manual batched scatter and gather kernels · d690c7b2
  Rick Ho authored Dec 30, 2020
  
  d690c7b2
- limit max streams · bf60a846
  Rick Ho authored Dec 30, 2020
  
  bf60a846
- naive memcpy batch with buffer · 9cae3294
  Rick Ho authored Dec 30, 2020
  
  9cae3294
- tide up C code · 8cff6ad7
  Rick Ho authored Dec 30, 2020
  
  8cff6ad7