Support converting arbitrary objects to StridedMemoryView in cuda.core.launch() #1280

Activity

1. leofang

   added this to the cuda.core backlog milestone on Nov 20, 2025
2. leofang

   added a parent issue on Nov 20, 2025

   • cuda.core.launch() improvements #150
3. github-actions

   added

   triageNeeds the team's attentionNeeds the team's attention

   on Nov 20, 2025
4. leofang commented on Nov 21, 2025

   

   leofang

   on Nov 21, 2025

   Last edited by leofang

   MemberAuthor

   More actions

   During the meeting today, the suggestion made by @kkraus14 was that we should insert a conversion to StridedMemoryView to the end of the loop in ParamHolder so that it is tried and this code

   launch(s, config, ker, a)

   is legal too, in addition to

   launch(s, config, ker, a.data.ptr)

   when a is a CuPy array. However, the exact type on the device side will have to be specified in this case (xref: #1281, #748).
5. leofang

   added

   enhancementAny code-related improvementsAny code-related improvements

   cuda.coreEverything related to the cuda.core moduleEverything related to the cuda.core module

   on Nov 24, 2025
6. leofang

   marked [FEA]: Autodetect __cuda_array_interface__ for launching kernels #1435 as a duplicate of this issue on Jan 7, 2026
7. rparolin

   removed

   triageNeeds the team's attentionNeeds the team's attention

   on Jan 19, 2026
8. rparolin

   removed this from the cuda.core backlog milestone on Apr 2, 2026
9. leofang commented on May 30, 2026

   

   leofang

   on May 30, 2026

   MemberAuthor

   More actions

   Please leave this task to us.