This repository has been archived by the owner on Sep 15, 2021. It is now read-only.

chore: initial baseline implementation #1

Merged

NOBLES5E merged 11 commits into master from baseline

Aug 19, 2021

Member

shjwudp commented Aug 13, 2021

No description provided.

shjwudp added 3 commits

August 4, 2021 21:46


          feat: stagging 1

eebef3d

..

ac15832


          feat: runable baseline

67663e1

shjwudp requested a review from NOBLES5E

August 13, 2021 03:43

todo bot commented Aug 13, 2021

support user specified interfaces

https://github.com/BaguaSys/bagua-nccl-net/blob/ac15832291f7c8f25d28ae8e357d38f615cd298e/src/utils.rs#L33-L38

This comment was generated by todo based on a `TODO` comment in `ac15832` in #1. cc @BaguaSys.

pr-triage bot added the PR: unreviewed label

..

3fd2771

todo bot commented Aug 13, 2021

fix this deadlock. tcp_stream must be nonblock, otherwise it may deadlock

https://github.com/BaguaSys/bagua-nccl-net/blob/3fd27713274e85cd8e48f55bf79f965085ebba00/src/bagua_net.rs#L151-L156

This comment was generated by todo based on a `TODO` comment in `3fd2771` in #1. cc @BaguaSys.

todo bot commented Aug 13, 2021

make isend true async

https://github.com/BaguaSys/bagua-nccl-net/blob/3fd27713274e85cd8e48f55bf79f965085ebba00/src/bagua_net.rs#L365-L370

This comment was generated by todo based on a `TODO` comment in `3fd2771` in #1. cc @BaguaSys.

todo bot commented Aug 13, 2021

make async read

https://github.com/BaguaSys/bagua-nccl-net/blob/3fd27713274e85cd8e48f55bf79f965085ebba00/src/bagua_net.rs#L445-L450

This comment was generated by todo based on a `TODO` comment in `3fd2771` in #1. cc @BaguaSys.

todo bot commented Aug 13, 2021

support parse sockaddr from NCCL_COMM_ID

https://github.com/BaguaSys/bagua-nccl-net/blob/3fd27713274e85cd8e48f55bf79f965085ebba00/src/utils.rs#L37-L42

This comment was generated by todo based on a `TODO` comment in `3fd2771` in #1. cc @BaguaSys.

..

5df9a7d

NOBLES5E reviewed

View reviewed changes

.gitignore Outdated Show resolved Hide resolved

NOBLES5E reviewed

View reviewed changes

README.md Outdated Show resolved Hide resolved

NOBLES5E reviewed

View reviewed changes

README.md Show resolved Hide resolved

NOBLES5E reviewed

View reviewed changes

cc/Makefile Outdated Show resolved Hide resolved

NOBLES5E reviewed

View reviewed changes

cc/bagua_net.cc Outdated Show resolved Hide resolved

NOBLES5E reviewed

View reviewed changes

src/bagua_net.rs Outdated Show resolved Hide resolved

NOBLES5E reviewed

View reviewed changes

Cargo.toml Show resolved Hide resolved

NOBLES5E suggested changes

View reviewed changes

Member

NOBLES5E left a comment

see comments

pr-triage bot added PR: reviewed-changes-requested and removed PR: unreviewed labels


          fix: resolve comments

d44bfc2

todo bot commented Aug 13, 2021

@shjwudp: support parse sockaddr from NCCL_COMM_ID

https://github.com/BaguaSys/bagua-nccl-net/blob/d44bfc2c89c9200f19d98daf94367fe3e486eb2c/src/utils.rs#L37-L42

This comment was generated by todo based on a `TODO` comment in `d44bfc2` in #1. cc @BaguaSys.

pr-triage bot added PR: unreviewed and removed PR: reviewed-changes-requested labels

shjwudp requested a review from NOBLES5E

August 13, 2021 06:59

shjwudp added 2 commits

August 16, 2021 15:03


          feat: Each tcp stream is allocated a thread

dfafe8d


          refactor: Refactor the communication model to configure communication…

7f50683

… threads according to tcp stream

todo bot commented Aug 18, 2021

make shutdown global

https://github.com/BaguaSys/bagua-nccl-net/blob/7f506837bfe9026f068c8befcc29068c8c8b6568/src/bagua_net.rs#L394-L399

This comment was generated by todo based on a `TODO` comment in `7f50683` in #1. cc @BaguaSys.


          docs: + benchmark data in README

dca7e6f

todo bot commented Aug 18, 2021

make Rotating communicator

https://github.com/BaguaSys/bagua-nccl-net/blob/dca7e6fe05757e41680c938840ee423ed8d4263f/src/bagua_net.rs#L57-L62

This comment was generated by todo based on a `TODO` comment in `dca7e6f` in #1. cc @BaguaSys.

shjwudp added 2 commits

August 18, 2021 16:58

..

0b9a1b2

..

4cb1bc9

Member Author

shjwudp commented Aug 18, 2021

@NOBLES5E The benchmark data has been updated to the README. Can we make the baseline become the master?

NOBLES5E changed the title ~~feat: runable baseline~~ chore: initial baseline implementation

NOBLES5E merged commit 9be4859 into master

NOBLES5E deleted the baseline branch

August 19, 2021 05:23

pr-triage bot added PR: merged and removed PR: unreviewed labels

This was referenced Aug 19, 2021

make Rotating communicator #2

Closed

Close opentelemetry at a more appropriate time #3

Open

@shjwudp: support parse sockaddr from NCCL_COMM_ID #4

Open

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.