Commit message (Expand) | Author | Age | |
---|---|---|---|
* | Dynamic subdivisions in collective ring reduce. | 2018-09-27 | |
* | Improve the GPU memory use discipline of CollectiveReduce. | 2018-08-29 | |
* | Refactor collectives to colocate implementation-specific code. | 2018-08-27 | |
* | In ring_reducer.cc if the reduction does not need a final op, then | 2018-08-02 | |
* | Unblock RingReducer's PCQueue after calling StartAbort in async callback. | 2018-07-02 | |
* | Add GPUOptions::num_dev_to_dev_copy_streams to allow creation of | 2018-06-27 | |
* | Merge changes from github. | 2018-05-24 | |
* | ring_reducer.cc errata: | 2018-05-10 | |
* | Delay deleting RingReducer until group_size_tensor_ready_ has | 2018-04-26 | |
* | Fix bug in ring_reducer.cc abort handling. | 2018-04-19 | |
* | Collective Ops Part 3 | 2018-04-12 |