-
Notifications
You must be signed in to change notification settings - Fork 916
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
Add initial support for string udfs in libcudf #10686
Conversation
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.
I'm uncomfortable with putting this class in libcudf. I had thought this was going to live somewhere else given that is solely for a numba feature, especially when it is a non-standard dstring
type thing that is kinda-sorta like std::string
but not all the way. imo, doing this "right" would be figuring out how to implement cuda::std::string
or cuda::string
in libcu++.
Requesting changes on this for now until we've had a chance to talk about this more.
Codecov Report
@@ Coverage Diff @@
## branch-22.06 #10686 +/- ##
================================================
+ Coverage 86.35% 86.38% +0.03%
================================================
Files 142 142
Lines 22335 22335
================================================
+ Hits 19287 19294 +7
+ Misses 3048 3041 -7
Continue to review full report at Codecov.
|
@brandon-b-miller and I discussed this a bit last week. We need the Anyway, we are reconsidering moving this into cudf right now in favor of trying to build a solution in the string_udf repo instead. |
Reference #9639
Add basic functions and a dynamic device string class in libcudf to support user-defined functions through numba. The device string class
dstring
manages local device memory to manipulate UTF-8 encoded string data in device code. The column functions help create and destroy an array ofdstring
objects andstring_view
objects for use within a numba created device kernel.A follow-on PR will include Cython interfaces for the column functions.
Depends on #10684