GitHub · Where software is built

Labels Milestones New issue

Accumulate becomes slow for very large input sizes

#75

· AntonReinhard opened

on Dec 17, 2025

Accumulate crashes on CUDA with DecoupledLookback

#74

· AntonReinhard opened

on Dec 17, 2025

`AcceleratedKernels.sum` is twice as slow than CUDA.jl's mapreduce implementation

#69

· VarLad opened

on Nov 12, 2025

Unexpected result from `accumulate`

#67

· jariji opened

on Nov 7, 2025

reducing while iterating over the index

#66

· albertomercurio opened

on Sep 21, 2025

Statistics kernels?

#62

· Moelf opened

on Aug 10, 2025

Use matrix input in `searchsortedfirst`

#61

· mtagliazucchi opened

on Aug 7, 2025

Use heuristic to determine block size instead of fixed value

#60

· christiangnrd opened

on Aug 5, 2025

Add `dims` support to GPU sorting algorithms

#59

· christiangnrd opened

on Aug 5, 2025

Support `mapreduce` with multiple source arrays

#54

· christiangnrd opened

on Jul 20, 2025

Support reducing on multiple dims

#53

· christiangnrd opened

on Jul 20, 2025

Performance comparison against KernelAbstractions.jl

#50

· mtagliazucchi opened

on Jul 5, 2025