Skip to content

Commit

Permalink
Considerable speedup(VGG model:1.5x, AlexNet:1.1x)
Browse files Browse the repository at this point in the history 
Optimizations focus on the gpu-related features, such as avoiding bank
conflict, employing wider band width of shared memory, and using
vectorized data type, etc..
  • Loading branch information
@bestimage-tencent
bestimage-tencent committed May 18, 2015
1 parent 074f352 commit e7739c8
Show file tree
Hide file tree
Showing 3 changed files with 841 additions and 394 deletions.
Loading

0 comments on commit e7739c8

Please sign in to comment.