Skip to content

SIMD acceleration using AVX2 #2

@WinPooh32

Description

@WinPooh32

SIMD implementation provided by viterin/vek

Build with float64 and avx2:

go build -tags "series_avx2"

Build with float32 and avx2:

go build -tags "series_f32 series_avx2"
Functon / Method Float32 Float64
Equals
IndexEquals
ValuesEquals
Add x x
Sub x x
Mul x x
Div x x
Mod
AddScalar x x
SubScalar x x
MulScalar x x
DivScalar x x
Sign
Sin x
Asin
Cos x
Acos
Tan
Atan
Pow
Pow10
Sqr
Exp x
Exp2
Log x
Log2 x
Log10 x
Abs
Floor x x
Trunc
Round x x
RoundToEven
Ceil x x
Cumsum
Reverse
IndexReverse
DataReverse
Fillna
Pad
Lerp
Diff
Shift

Aggregation:

Functon / Method Float32 Float64
Sum
Min
Max
Skew
Variance
Std
First
Last
Argmin
Argmax
Median

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions