-
Notifications
You must be signed in to change notification settings - Fork 448
Conversation
For both Device and Block algorithms. Note that these do not handle twiddling during descending sort in the same way.
@zasdfgbnm I added some more documentation to the sort algorithms in the last push, including #293. Can you review these changes to make sure they provide enough information for your needs? |
* \par Supported Types | ||
* BlockRadixSort can sort all of the built-in C++ numeric primitive types | ||
* (<tt>unsigned char</tt>, \p int, \p double, etc.) as well as CUDA's \p __half | ||
* half-precision floating-point type. |
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 might need to add __nv_bfloat16
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 have updated #306 to mention bfloat16. I'll update that PR after this goes in.
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 change of doc is very good. I like it. I left some comments. @ngimel do you have any further comments?
Fixes #282, extends and replaces #284.