Re:Multiplatform Matrix Multiplication Kernels
by nathanielsimard on Friday July 18, 2025 @06:48PM
(#44610643)
Well we can agree to disagree, CubeCL also has the concept of instruction parallelism, which would be used to target simd instructions on CPU. Our algorithms are normally flexible on both the plane size and the line size, adapting to the hardware with comptime logique. You are free to dislike the naming, but imo a mix of multiple APIs is worse than something new.
[ Parent
| on HN ]
Re:Multiplatform Matrix Multiplication Kernels
by gyrovagueGeist on Friday July 18, 2025 @07:18PM
(#44610891)
For people who are interested Kokkos (a C++ library for writing portable kernels) also has a naming scheme for hierarchical parallelism. They use ThreadTeam, Thread (for individual threads within a group), and ThreadVector (for per thread SIMD).
Just commenting to share, personally I have no naming preference but the hierarchal abstractions in general are incredibly useful.
Re:Multiplatform Matrix Multiplication Kernels
by almostgotcaught on Friday July 18, 2025 @06:53PM
(#44610670)
> Our algorithms are normally flexible on both the plane size and the line size
Congrats - I have no idea what this means lol.
Re:Multiplatform Matrix Multiplication Kernels
by syl20bnr on Friday July 18, 2025 @07:44PM
(#44611092)
It will make more sense once you start using CubeCL. There's now a CubeCL book available: https://burn.dev/books/cubecl/.
It does come with some mental overhead, but let’s be honest, there’s no objectively “good” choice here without introducing bias toward a specific vendor API.
Learning the core concepts takes effort, but if CubeCL is useful for your work, it’s definitely worth it.