Commits · 96ee9571e2c96ba6eb6972da1be75453d6c6e9fa · yangql / composable_kernel

"...resnet50_tensorflow.git" did not exist on "f27ec7f61226224efd03d4bcce306521ba7067f0"

10 Apr, 2019 6 commits
- tuned implicit gemm v1 for 3x3 on AMD to 82%. Fixed a bug in 4d tensor blockwise copy. · 96ee9571
  Chao Liu authored Apr 10, 2019
  
  96ee9571
- update flops calculation · edc89778
  Chao Liu authored Apr 10, 2019
  
  edc89778
- simplify blockwise batched GEMM · 5696c81f
  Chao Liu authored Apr 10, 2019
  
  5696c81f
- add 3x3 28x28 case · 71434918
  Chao Liu authored Apr 10, 2019
  
  71434918
- clean up · d86a5e4b
  Chao Liu authored Apr 09, 2019
  
  d86a5e4b
- enabled ds_read_b128 and ds_write_b128 on hip c++ · e624df92
  Chao Liu authored Apr 09, 2019
  
  e624df92
09 Apr, 2019 3 commits
- tidy yp · 471830a0
  Chao Liu authored Apr 09, 2019
  
  471830a0
- refactor · 1bd880a6
  Chao Liu authored Apr 09, 2019
  
  1bd880a6
- load smaller weight tensor · 796f72e2
  Chao Liu authored Apr 08, 2019
  
  796f72e2
08 Apr, 2019 5 commits
- refactor · 5b36aead
  Chao Liu authored Apr 08, 2019
  
  5b36aead
- added implicit_gemm_v1 lds double_buffer · cc0fa73a
  Chao Liu authored Apr 08, 2019
  
  cc0fa73a
- add more assertion · c075d3f7
  Chao Liu authored Apr 08, 2019
  
  c075d3f7
- tidy up · 268d1c71
  Chao Liu authored Apr 08, 2019
  
  268d1c71
- debugging implicit gemm v1: use 10d tensor output · c9fa46af
  Chao Liu authored Apr 08, 2019
  
  c9fa46af
07 Apr, 2019 2 commits
- refactor · 90abf427
  Chao Liu authored Apr 06, 2019
  
  90abf427
- refactor · b57d60c0
  Chao Liu authored Apr 06, 2019
  
  b57d60c0
06 Apr, 2019 4 commits
- use dedicated threadwise_copy for 1x1, perf at 80% · bd0098af
  Chao Liu authored Apr 06, 2019
  
  bd0098af
- clean up · 5245a016
  Chao Liu authored Apr 06, 2019
  
  5245a016
- debugged: CUDA should use its own float4 definition · 7a251a09
  Chao Liu authored Apr 06, 2019
  
  7a251a09
- debugging · f6cb5b84
  Chao Liu authored Apr 06, 2019
  
  f6cb5b84
05 Apr, 2019 3 commits
- debugging · 0983d205
  Chao Liu authored Apr 05, 2019
  
  0983d205
- preload don't wait · bae23337
  Chao Liu authored Apr 05, 2019
  
  bae23337
- clipboard float4 copy and paste C++ code · dabfa77f
  Chao Liu authored Apr 05, 2019
  
  dabfa77f
04 Apr, 2019 9 commits
- Merge branch 'master' into inline_asm_v2 · 605afd0f
  Chao Liu authored Apr 04, 2019
  
  605afd0f
- Merge branch 'inline_asm_v2' of github.com:asroy/modular_convolution into inline_asm_v2 · 66edb259
  Chao Liu authored Apr 04, 2019
  
  66edb259
- unroll even-odd loop · 19b41797
  Chao Liu authored Apr 04, 2019
  
  19b41797
- clean code · 62c4d5df
  Jing Zhang authored Apr 04, 2019
  
  62c4d5df
- unroll k · 313f3c07
  Jing Zhang authored Apr 04, 2019
  
  313f3c07
- add debug · 0f620a90
  Jing Zhang authored Apr 04, 2019
  
  0f620a90
- add asm into lds_double_buffer version · fbc7817b
  Chao Liu authored Apr 04, 2019
  
  fbc7817b
- optimize global_load · 155d7859
  Jing Zhang authored Apr 03, 2019
  
  155d7859
- enable 128x128 block gemm · 05d7a087
  Chao Liu authored Apr 03, 2019
  
  05d7a087
03 Apr, 2019 8 commits
- add · 6a3f3f95
  Jing Zhang authored Apr 03, 2019
  
  6a3f3f95
- fixed wait · b188c0d2
  Jing Zhang authored Apr 03, 2019
  
  b188c0d2
- increase depth of pipeline · 0de4286a
  Jing Zhang authored Apr 03, 2019
  
  0de4286a
- add inline asm · 6fef303e
  Jing Zhang authored Apr 03, 2019
  
  6fef303e
- inline asm · 0d6aa311
  Jing Zhang authored Apr 03, 2019
  
  0d6aa311
- refactor inline asm · 753b98b5
  Jing Zhang authored Apr 03, 2019
  
  753b98b5
- add script to extrac asm on hip · 6166233e
  Chao Liu authored Apr 03, 2019
  
  6166233e
- tidy up · e2313c9e
  Chao Liu authored Apr 02, 2019
  
  e2313c9e