-
-
Notifications
You must be signed in to change notification settings - Fork 0
Open
Labels
FeatureTask to add a new feature.Task to add a new feature.difficulty: 2May require some initial design or R&D, but should be straightforward to resolve and/or implement.May require some initial design or R&D, but should be straightforward to resolve and/or implement.
Description
This API is similar to the Cartesian product and the following APIs found elsewhere
- https://docs.pytorch.org/docs/stable/generated/torch.combinations.html
- https://www.mathworks.com/help/matlab/ref/nchoosek.html
API signature:
function dcombinations( N, k, replacement, x, strideX, out, strideOut1, strideOut2 )where out is a 2-dimensional strided array which is assumed to hold at least N choose k elements if replacement is false and N+k-1 choose k elements if replacement is true.
This (and similar (s|z|c|g) APIs will set the stage for an eventual blas/ext/combinations API.
Because out is two-dimensional, the implementation should be cache-aware (i.e., set elements in out in a cache-optimal fashion).
Metadata
Metadata
Assignees
Labels
FeatureTask to add a new feature.Task to add a new feature.difficulty: 2May require some initial design or R&D, but should be straightforward to resolve and/or implement.May require some initial design or R&D, but should be straightforward to resolve and/or implement.