You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently Blosc.jl allows for compression of DenseArrays that are contiguous. Besides being contiguous, the main requirement for an AbstractArray is the applicability of pointer.
Base.FastContiguousSubArray satisfies contiguity and has method pointer.
julia> Base.FastContiguousSubArray
SubArray{T,N,P,I,true} where I<:Union{Tuple{Vararg{Real,N} where N}, Tuple{AbstractUnitRange,Vararg{Any,N} where N}}where P where N where T
julia>hasmethod(pointer, (Base.FastContiguousSubArray,))
true
SIMD.FastContiguousArray also satisfies contiguity and has method pointer
julia>using SIMD
julia> SIMD.FastContiguousArray
Union{DenseArray{T,N}, Base.ReinterpretArray{T,N,T2,A} where A<:Union{DenseArray{T2,N}, SubArray{T2,N,P,I,true} where I<:Union{Tuple{Vararg{Real,N} where N}, Tuple{AbstractUnitRange,Vararg{Any,N} where N}}where P} where T2, SubArray{T,N,P,I,true} where I<:Union{Tuple{Vararg{Real,N} where N}, Tuple{AbstractUnitRange,Vararg{Any,N} where N}}where P} where N where T
julia>hasmethod(pointer, (SIMD.FastContiguousArray,))
true
Currently Blosc.jl allows for compression of
DenseArray
s that are contiguous. Besides being contiguous, the main requirement for anAbstractArray
is the applicability ofpointer
.Base.FastContiguousSubArray
satisfies contiguity and has methodpointer
.SIMD.FastContiguousArray
also satisfies contiguity and has methodpointer
https://github.com/eschnett/SIMD.jl/blob/master/src/arrayops.jl#L40-L42
Could we widen the
compress
,compress!
,decompress
, anddecompress!
methods to include these types, perhaps in place ofDenseArray
?The text was updated successfully, but these errors were encountered: