OP primitive cache: use memory as signature for MKLDNN storage type #11
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
5 tasks
|
@TaoLv the case you mentioned doesn't happen. IsMKLDNN returns true, the NDArray must have mkldnn memory attached to it. |
zheng-da
added a commit
that referenced
this pull request
Jun 1, 2018
* add functions for cutting edges. * construct subgraphs. * generalize graph partition. * restructure the code. * create SubgraphOpState. * register subgraph property. * rename. * address comments. * update select API. * rename. * set subgraph property. * fix bugs. * fix bugs.
zheng-da
added a commit
that referenced
this pull request
Jun 7, 2018
* add functions for cutting edges. * construct subgraphs. * generalize graph partition. * restructure the code. * create SubgraphOpState. * register subgraph property. * rename. * address comments. * update select API. * rename. * set subgraph property. * fix bugs. * fix bugs.
zheng-da
added a commit
that referenced
this pull request
Jun 13, 2018
* add functions for cutting edges. * construct subgraphs. * generalize graph partition. * restructure the code. * create SubgraphOpState. * register subgraph property. * rename. * address comments. * update select API. * rename. * set subgraph property. * fix bugs. * fix bugs.
TaoLv
pushed a commit
to TaoLv/incubator-mxnet
that referenced
this pull request
Jun 16, 2018
* add functions for cutting edges. * construct subgraphs. * generalize graph partition. * restructure the code. * create SubgraphOpState. * register subgraph property. * rename. * address comments. * update select API. * rename. * set subgraph property. * fix bugs. * fix bugs.
TaoLv
pushed a commit
to TaoLv/incubator-mxnet
that referenced
this pull request
Feb 18, 2019
Opt 1.4.0rc2
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
For Convolution/FC OPs, we can use NDArray dims/shape to do primitive cache signature, for some other OPs, we need to consider memory layout also, to unify the interface, we would like all OPs to pass NDArray type to AddSign and AddSign function will convert NDArray to memory type and do signature. With this change, convolution/FC OPs will also use memory to do signature, from testing, we almost didn't see any increase for convolution primitive cache count (using NDArray vs using memory to do sign) so it probably should be ok to use memory to sign convolution also.
Checklist
Essentials
make lint)