-
Notifications
You must be signed in to change notification settings - Fork 19
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
Distributed FMM with complete tree structure on each rank #8
Distributed FMM with complete tree structure on each rank #8
Conversation
…tree into distributed-fmm-global
ddb2065
to
5d13e04
Compare
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.
Hi Hao, thanks for working on this! Here's a bit more progress on review.
FYI, the "tree-dependent data" PRs:
- Separate kernel-dependent/tree-independent state out of wrangler #29
- Track TreeIndependentData refactor from boxtree, remove persistent queue from wrangler sumpy#70
- Track tree_indep refactor in boxtree and sumpy pytential#101
are done and should be landing soon.
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.
(as discussed during meeting)
boxtree/distributed/__init__.py
Outdated
calibration_params = \ | ||
FMMCostModel.get_unit_calibration_params() | ||
|
||
# We need to construct a wrangler in order to access `level_nterms` |
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.
This reverts commit 8b2e392.
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.
Unbelievable. I actually remembered to submit comments this time.
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.
Thanks! A few more comments below, largely minor. In terms of review progress, I've now read everything except
local_traversal
local_tree
partition
traversal
FYI I just rebased inducer/sumpy#70 to bring it up to date with respect to the latest sumpy (actually just redid the patch, the merge didn't really work). |
Whoops! I did not mean to close this. I think what happened here is that I merged the wrangler refactor (#29), which served as the base branch of this, and because that branch got deleted, it closed this PR. @gaohao95, could you open a new PR off of your branch that targets |
TODO List:
Must do:
test_distributed.py
.calculate_pot
for Sumpy.Maybe:
box_mpole_is_used
.