-
Notifications
You must be signed in to change notification settings - Fork 3.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Heterogeneous][Bugfix] Fix bug of wrongly generated device_map #2990
Conversation
@imorinaga Thanks for the fix. It looks good to me. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
otherwise good to me
add mul | ||
\ / | ||
subtract | ||
device_index is wrongly propagated in previous version""" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
remove this comment?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
done
Thanks @imorinaga @zhiics This is merged. |
…he#2990) * fix bug of device_index * cpplint * nose * Update test_pass_annotation.py * fix name of testcase * delete comment
…he#2990) * fix bug of device_index * cpplint * nose * Update test_pass_annotation.py * fix name of testcase * delete comment
For the following type of graph, device_index should be [2, 1, 2, 2] (corresponding to [a, b, copy1, c]) but generated one is [1, 1, 2, 2].
The problems are
This patch fix these by checking whether each nodes is bottom-up propagatable when creating post_dfs_order_.
@zhiics could you please review?