Inline most functions

Sources/AccelerateLinux/MatrixOps/MatrixBasicOps.swift

@@ -12,6 +12,8 @@ import CBLAS
 ///   - __M: The number of rows in the output matrix and the number of columns in the input matrix.
 ///   - __N: The number of columns in the output matrix and the number of rows in the input matrix.
 /// - Warning: This function doesn’t support in-place operation.
+@inlinable
+@inline(__always)
 public func vDSP_mtransD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -42,6 +44,8 @@ public func vDSP_mtransD(
 ///   - __M: The number of rows in matrices A and C.
 ///   - __N: The number of columns in matrices B and C.
 ///   - __P: The number of columns in matrix A and the number of rows in matrix B.
+@inlinable
+@inline(__always)
 public func vDSP_mmulD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,

Sources/AccelerateLinux/VectorOps/VectorBasicOps.swift

@@ -10,6 +10,8 @@ import Glibc
 ///  - __I: The stride of the vector.
 ///  - __C: The maximum value.
 ///  - __N: The number of elements in the vector.
+@inlinable
+@inline(__always)
 public func vDSP_maxvD(
     _ __A: UnsafePointer<Double>,
     _ __I: vDSP_Stride,
@@ -31,6 +33,8 @@ public func vDSP_maxvD(
 ///   - __I: The stride of the vector.
 ///   - __C: The minimum value.
 ///   - __N: The number of elements in the vector.
+@inlinable
+@inline(__always)
 public func vDSP_minvD(
     _ __A: UnsafePointer<Double>,
     _ __I: vDSP_Stride,
@@ -55,6 +59,8 @@ public func vDSP_minvD(
 ///   - __C: The output vector, C.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vaddD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -80,6 +86,8 @@ public func vDSP_vaddD(
 ///   - __C: The output vector, C.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vsubD(
     _ __B: UnsafePointer<Double>,
     _ __IB: vDSP_Stride,
@@ -105,6 +113,8 @@ public func vDSP_vsubD(
 ///   - __C: The output vector, C.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vmulD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -127,6 +137,8 @@ public func vDSP_vmulD(
 ///   - _: The exponent input array, y.
 ///   - _: The base input array, x.
 ///   - _: The number of elements in the arrays.
+@inlinable
+@inline(__always)
 public func vvpow(
     _ z: UnsafeMutablePointer<Double>,
     _ y: UnsafePointer<Double>,
@@ -158,6 +170,8 @@ public func vvpow(
 ///   - __C: Double-precision real output vector.
 ///   - __IC: Address stride for C.
 ///   - __N: The number of elements to process.
+@inlinable
+@inline(__always)
 public func vDSP_vclrD(
     _ __C: UnsafeMutablePointer<Double>,
     _ __IC: vDSP_Stride,
@@ -188,6 +202,8 @@ public func vDSP_vfillD(
 ///   - __C: On output, the absolute values of the elements in the input vector.
 ///   - __IC: The distance between the elements in the output vector C.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vabsD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -209,6 +225,8 @@ public func vDSP_vabsD(
 ///   - __C: On output, the negative values of the elements in the input vector.
 ///   - __IC: The distance between the elements in the output vector C.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vnegD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -224,6 +242,8 @@ public func vDSP_vnegD(
 }
 
 /// Computes the squared value of each element in the supplied double-precision vector.
+@inlinable
+@inline(__always)
 public func vDSP_vsqD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -246,6 +266,8 @@ public func vDSP_vsqD(
 ///   - __IB: The distance between the elements in the input vector B.
 ///   - __C: On output, the dot product of the two vectors.
 ///   - __N: The number of elements to process.
+@inlinable
+@inline(__always)
 public func vDSP_dotprD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -263,6 +285,8 @@ public func vDSP_dotprD(
 }
 
 /// Calculates the double-precision vector test limit using the specified stride.
+@inlinable
+@inline(__always)
 public func vDSP_vlimD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -290,6 +314,8 @@ public func vDSP_vlimD(
 ///   - __N: The number of elements that the function clips to the specified range.
 ///   - __NLow: On output, the number of elements clipped to the low-clipping threshold.
 ///   - __NHigh: On output, the number of elements clipped to the high-clipping threshold.
+@inlinable
+@inline(__always)
 public func vDSP_vclipcD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -327,6 +353,8 @@ public func vDSP_vclipcD(
 /// Integrates vector A using a running sum from vector C.
 /// Vector A is weighted by scalar *S and added to the previous output point.
 /// The first element from vector A is not used in the sum.
+@inlinable
+@inline(__always)
 public func vDSP_vrsumD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -356,6 +384,8 @@ public func vDSP_vrsumD(
 ///   - __N: The number of elements in the vector.
 ///   - __Order: A value that specifies the sort order. Pass 1 to specify ascending order, or -1 for descending order.
 #warning("If the array is bigger than Int.max, vDSP_vsortD will do nothing. Find a solution")
+@inlinable
+@inline(__always)
 public func vDSP_vsortD(
     _ __C: UnsafeMutablePointer<Double>,
     _ __N: vDSP_Length,
@@ -388,6 +418,8 @@ public func vDSP_vsortD(
 ///   - __C: The output vector.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vrampD(
     _ __A: UnsafePointer<Double>,
     _ __B: UnsafePointer<Double>,
@@ -416,6 +448,8 @@ public func vDSP_vrampD(
 /// for (n = 0; n < N; ++n)
 ///    C[n] = A[n] <= B[n] ? A[n] : B[n];
 /// ```
+@inlinable
+@inline(__always)
 public func vDSP_vminD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -441,6 +475,8 @@ public func vDSP_vminD(
 ///   - __C: Double-precision real output vector.
 ///   - __IC: Stride for C.
 ///   - __N: The number of elements to process.
+@inlinable
+@inline(__always)
 public func vDSP_vmaxD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -466,6 +502,8 @@ public func vDSP_vmaxD(
 ///   - __C: The output vector, C.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vdivD(
     _ __B: UnsafePointer<Double>,
     _ __IB: vDSP_Stride,
@@ -490,6 +528,8 @@ public func vDSP_vdivD(
 ///   - __C: The output vector, C.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vsdivD(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -513,6 +553,8 @@ public func vDSP_vsdivD(
 ///
 /// If x is +/-0, the result preserves the signed zero.
 /// If x is +/-inf, the result is NaN.
+@inlinable
+@inline(__always)
 public func vvsin(
     _ y: UnsafeMutablePointer<Double>,
     _ x: UnsafePointer<Double>,
@@ -530,6 +572,8 @@ public func vvsin(
 ///   - y: The output array, y.
 ///   - x: The input array, x.
 ///   - n: The number of elements in the arrays.
+@inlinable
+@inline(__always)
 public func vvsqrt(
     _ y: UnsafeMutablePointer<Double>,
     _ x: UnsafePointer<Double>,
@@ -547,6 +591,8 @@ public func vvsqrt(
 ///   - y: The output array, y.
 ///   - x: The input array, x.
 ///   - n: The number of elements in the arrays.
+@inlinable
+@inline(__always)
 public func vvexp(
     _ y: UnsafeMutablePointer<Double>,
     _ x: UnsafePointer<Double>,
@@ -564,6 +610,8 @@ public func vvexp(
 ///   - y: The output array, y.
 ///   - x: The input array, x.
 ///   - n: The number of elements in the arrays.
+@inlinable
+@inline(__always)
 public func vvlog(
     _ y: UnsafeMutablePointer<Double>,
     _ x: UnsafePointer<Double>,
@@ -576,7 +624,8 @@ public func vvlog(
     }
 }
 
-private func quicksort(
+@usableFromInline
+func quicksort(
     _ vec: UnsafeMutablePointer<Double>,
     _ len: vDSP_Length,
     _ ord: Int32,

Sources/AccelerateLinux/VectorOps/VectorTransformOps.swift

@@ -9,6 +9,8 @@
 ///   - __C: The output vector.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vdpsp(
     _ __A: UnsafePointer<Double>,
     _ __IA: vDSP_Stride,
@@ -30,6 +32,8 @@ public func vDSP_vdpsp(
 ///   - __C: The output vector.
 ///   - __IC: The distance between the elements in the output vector.
 ///   - __N: The number of elements that the function processes.
+@inlinable
+@inline(__always)
 public func vDSP_vspdp(
     _ __A: UnsafePointer<Float>,
     _ __IA: vDSP_Stride,

Sources/AccelerateLinux/VectorOps/vDSP+StatisticalOps.swift

@@ -5,6 +5,8 @@
 extension vDSP {
     /// Returns the mean value of a double-precision vector.
     /// - Parameter vector: The source vector.
+    @inlinable
+    @inline(__always)
     public static func mean<U>(_ vector: U) -> Double where U: AccelerateBuffer, U.Element == Double {
         sum(vector) / Double(vector.count)
     }