Core OS #WWDC14

What’s New in the 


Accelerate Framework

Session 703
Geoff Belter
Engineer, Vector and Numerics Group

© 2014 Apple Inc. All rights reserved. Redistribution or public display not permitted without written permission from Apple.
What is Available?

Image processing (vImage)

Digital signal processing (vDSP)
Math functions (vForce, vMathLib, vBigNum)
Linear algebra (LAPACK, BLAS)
What is the Accelerate Framework?

High performance
• Fast
• Energy efficient
OS X and iOS
All generations of hardware
Session Goals

New features in vImage

Introduce
• LinearAlgebra
• <simd/simd.h>
vImage
High performance image processing
vImage
Some things you can do
vImage
Some things you can do
Getting Data into vImage
CGImageRef
Getting Data into vImage
CGImageRef

// CGImageRef —-> vImage_Buffer


vImage_Buffer buf;

vImage_CGImageFormat fmt = { .bitsPerComponent = 8, .bitsPerPixel = 32, … };

vImage_Error err = vImageBuffer_initWithCGImage( &buf, &fmt, NULL, 

cgImage, kvImageNoFlags );



Getting Data into vImage
CGImageRef

// CGImageRef —-> vImage_Buffer


vImage_Buffer buf;

vImage_CGImageFormat fmt = { .bitsPerComponent = 8, .bitsPerPixel = 32, … };

vImage_Error err = vImageBuffer_initWithCGImage( &buf, &fmt, NULL, 

cgImage, kvImageNoFlags );




// vImage_Buffer —-> CGImageRef


cgImage = vImageCreateCGImageFromBuffer( &buf, &fmt, NULL, NULL,
kvImageNoFlags, &err );
Conversion Support
vImageConvert_AnyToAny

// 1) Make converter: srcFormat —-> destFormat


vImage_CGImageFormat srcFormat = { .bitsPerComponent = 8, … };

vImage_CGImageFormat destFormat = { .bitsPerComponent = 16, … };


vImageConverterRef c = vImageConverter_CreateWithCGImageFormat(

&srcFormat, &destFormat, NULL, kvImageNoFlags, &err);
!

// 2) Convert

vImage_Buffer srcBuf = {…}, destBuf = {…};

err = vImageConvert_AnyToAny(c, &srcBuf, &destBuf, NULL, flags);

Conversion Support
vImageConvert_AnyToAny

// 1) Make converter: srcFormat —-> destFormat


vImage_CGImageFormat srcFormat = { .bitsPerComponent = 8, … };

vImage_CGImageFormat destFormat = { .bitsPerComponent = 16, … };


vImageConverterRef c = vImageConverter_CreateWithCGImageFormat(

&srcFormat, &destFormat, NULL, kvImageNoFlags, &err);
!

// 2) Convert

vImage_Buffer srcBuf = {…}, destBuf = {…};

err = vImageConvert_AnyToAny(c, &srcBuf, &destBuf, NULL, flags);

Conversion Support
vImageConvert_AnyToAny

// 1) Make converter: srcFormat —-> destFormat


vImage_CGImageFormat srcFormat = { .bitsPerComponent = 8, … };

vImage_CGImageFormat destFormat = { .bitsPerComponent = 16, … };


vImageConverterRef c = vImageConverter_CreateWithCGImageFormat(

&srcFormat, &destFormat, NULL, kvImageNoFlags, &err);
!

// 2) Convert

vImage_Buffer srcBuf = {…}, destBuf = {…};

err = vImageConvert_AnyToAny(c, &srcBuf, &destBuf, NULL, flags);

Conversion Support
vImageConvert_AnyToAny

// 1) Make converter: srcFormat —-> destFormat


vImage_CGImageFormat srcFormat = { .bitsPerComponent = 8, … };

vImage_CGImageFormat destFormat = { .bitsPerComponent = 16, … };


vImageConverterRef c = vImageConverter_CreateWithCGImageFormat(

&srcFormat, &destFormat, NULL, kvImageNoFlags, &err);
!

// 2) Convert

vImage_Buffer srcBuf = {…}, destBuf = {…};

err = vImageConvert_AnyToAny(c, &srcBuf, &destBuf, NULL, flags);

 What You Had to Say

“functions that convert

vImage_Buffer objects to
CGImageRef objects and back 

👍👍👍👍👍👍”
!

Twitter user
 What You Had to Say

“functions that convert “vImageConvert_AnyToAny 


vImage_Buffer objects to is magical. Threaded and
CGImageRef objects and back 
 vectorized conversion between
👍👍👍👍👍👍” nearly any two pixel formats.”
!

Twitter user Twitter user

Video—RGB, Grayscale, and Y’CbCr
CVPixelBufferRef (a video frame)
Video—RGB, Grayscale, and Y’CbCr
CVPixelBufferRef (a video frame)

// CVPixelBufferRef —-> vImageBuffer


vImageBuffer_InitWithCVPixelBuffer( &buf, &desiredFormat, cvPixelBuffer,

NULL, NULL, kvImageNoFlags );
Video—RGB, Grayscale, and Y’CbCr
CVPixelBufferRef (a video frame)

// CVPixelBufferRef —-> vImageBuffer


vImageBuffer_InitWithCVPixelBuffer( &buf, &desiredFormat, cvPixelBuffer,

NULL, NULL, kvImageNoFlags );

// vImageBuffer —-> CVPixelBufferRef


vImageBuffer_CopyToCVPixelBuffer( &buf, &bufFormat, cvPixelBuffer, 

NULL, NULL, kvImageNoFlags );
Getting video into vImage

Lower level interfaces

• 41 video conversions
• Manage chroma siting, transfer function, conversion matrix, etc.
• RGB colorspaces for video formats


vImageConvert_AnyToAny() for video

- vImageConverter_CreateForCGtoCVImageFormat
- vImageConverter_CreateForCVtoCGImageFormat
VideoToolbox Performance
Higher is better
3000 OS X 10.9
OS X 10.10 (with vImage)

2250
MPixels/second

2035
1869
1739
1500

1219

750

458 460
344
0
420v yuvs y420 v210
LinearAlgebra (LA)
Simple to use high performance
Solving System of Linear Equations
With LAPACK

Given the system A, and the right-hand-side B, how do you find X (AX = B)?
Solving System of Linear Equations
With LAPACK

Given the system A, and the right-hand-side B, how do you find X (AX = B)?
!

__CLPK_integer n = matrix_size;
__CLPK_integer nrhs = number_right_hand_sides;
__CLPK_integer lda = column_stride_A;
__CLPK_integer ldb = column_stride_B;
__CLPK_integer *ipiv = malloc(sizeof(__CLPK_integer)*n);
__CLPK_integer info;
sgesv_(&n, &nrhs, A, &lda, ipiv, B, &ldb, &info);
free(ipiv);
Solving System of Linear Equations
With LA

Given the system A, and the right-hand-side B, how do you find X (AX = B)?
Solving System of Linear Equations
With LA

Given the system A, and the right-hand-side B, how do you find X (AX = B)?

la_object_t X = la_solve(A,B);
LinearAlgebra

New in iOS 8.0 and OS X Yosemite

LinearAlgebra

New in iOS 8.0 and OS X Yosemite

Simple with good performance
LinearAlgebra

New in iOS 8.0 and OS X Yosemite

Simple with good performance
Single and double precision
LinearAlgebra

New in iOS 8.0 and OS X Yosemite

Simple with good performance
Single and double precision
Native Objective-C Object
What’s Available?

Element-wise arithmetic
Matrix product
Transpose
Norms / normalization
Linear systems
Slice
Splat
LA Objects
LA Objects

Reference counted opaque objects

• Objective-C objects when appropriate
LA Objects

Reference counted opaque objects

• Objective-C objects when appropriate
Managed for you
• Data buffer/memory
• Dimension details
• Errors and warnings
• Scalar data type (float/double)
Memory Management
Memory Management

la_release()
la_retain()

la_object_t A, B, C;
// create A and B
C = la_sum(A,B);
la_release(A);
la_release(B);
// use C
la_release(C);
Memory Management

C Objective-C no ARC

la_release() -[release]
la_retain() -[retain]

la_object_t A, B, C; la_object_t A, B, C;
// create A and B // create A and B
C = la_sum(A,B); C = la_sum(A,B);
la_release(A); [A release];
la_release(B); [B release];
// use C // use C
la_release(C); [C release];
Memory Management

C Objective-C no ARC Objective-C with ARC

la_release() -[release]
la_retain() -[retain]

la_object_t A, B, C; la_object_t A, B, C; la_object_t A, B, C;

// create A and B // create A and B // create A and B
C = la_sum(A,B); C = la_sum(A,B); C = la_sum(A,B);
la_release(A); [A release]; !
la_release(B); [B release]; !
// use C // use C // use C
la_release(C); [C release];
Buffer to LA Object
With copy

double *A = malloc(sizeof(double) * num_rows * row_stride);

// Fill A as row major matrix
!

// Data copied into object, user still responsible for A

la_object_t Aobj = la_matrix_from_double_buffer(A, num_rows, num_cols,
row_stride, LA_NO_HINT,
LA_DEFAULT_ATTRIBUTES);
!

// User retains all rights to A. User must clean up A

free(A);
Buffer to LA Object
With copy

double *A = malloc(sizeof(double) * num_rows * row_stride);

// Fill A as row major matrix
!

// Data copied into object, user still responsible for A

la_object_t Aobj = la_matrix_from_double_buffer(A, num_rows, num_cols,
row_stride, LA_NO_HINT,
LA_DEFAULT_ATTRIBUTES);
!

// User retains all rights to A. User must clean up A

free(A);
Buffer to LA Object
With copy

double *A = malloc(sizeof(double) * num_rows * row_stride);

// Fill A as row major matrix
!

// Data copied into object, user still responsible for A

la_object_t Aobj = la_matrix_from_double_buffer(A, num_rows, num_cols,
row_stride, LA_NO_HINT,
LA_DEFAULT_ATTRIBUTES);
!

// User retains all rights to A. User must clean up A

free(A);
Buffer to LA Object
With copy

double *A = malloc(sizeof(double) * num_rows * row_stride);

// Fill A as row major matrix
!

// Data copied into object, user still responsible for A

la_object_t Aobj = la_matrix_from_double_buffer(A, num_rows, num_cols,
row_stride, LA_NO_HINT,
LA_DEFAULT_ATTRIBUTES);
!

// User retains all rights to A. User must clean up A

free(A);
Hints

la_object_t o = la_matrix_from_double_buffer(A, num_rows, num_cols,

row_stride, LA_SHAPE_DIAGONAL,
LA_DEFAULT_ATTRIBUTES);
Hints

la_object_t o = la_matrix_from_double_buffer(A, num_rows, num_cols,

row_stride, LA_SHAPE_DIAGONAL,
LA_DEFAULT_ATTRIBUTES);
Allow for better performance
Hints

la_object_t o = la_matrix_from_double_buffer(A, num_rows, num_cols,

row_stride, LA_SHAPE_DIAGONAL,
LA_DEFAULT_ATTRIBUTES);
Allow for better performance
Insight about the data buffer
• Diagonal
• Triangular
• Symmetric
• Positive Definite
Lazy Evaluation

la_object_t foo(la_object_t A, la_object_t x) {

// At = A’
la_object_t At = la_transpose(A);
!

// sum odd elements of x to even elements of x

la_object_t x2 = la_sum(la_vector_slice(x,0,2,la_vector_length(x)/2),
la_vector_slice(x,1,2,la_vector_length(x)/2));
!

// Atx2 = A’ * x2 * 3.2
la_object_t Atx2 = la_scale_with_float(la_matrix_product(At,x2), 3.2f);
if (la_status(Atx2) < 0) { // error }
return Atx2
}
Lazy Evaluation A x

la_object_t foo(la_object_t A, la_object_t x) {

// At = A’
la_object_t At = la_transpose(A);
!

// sum odd elements of x to even elements of x

la_object_t x2 = la_sum(la_vector_slice(x,0,2,la_vector_length(x)/2),
la_vector_slice(x,1,2,la_vector_length(x)/2));
!

// Atx2 = A’ * x2 * 3.2
la_object_t Atx2 = la_scale_with_float(la_matrix_product(At,x2), 3.2f);
if (la_status(Atx2) < 0) { // error }
return Atx2
}
Lazy Evaluation A x

At

la_object_t foo(la_object_t A, la_object_t x) {

// At = A’
la_object_t At = la_transpose(A);
!

// sum odd elements of x to even elements of x

la_object_t x2 = la_sum(la_vector_slice(x,0,2,la_vector_length(x)/2),
la_vector_slice(x,1,2,la_vector_length(x)/2));
!

// Atx2 = A’ * x2 * 3.2
la_object_t Atx2 = la_scale_with_float(la_matrix_product(At,x2), 3.2f);
if (la_status(Atx2) < 0) { // error }
return Atx2
}
Lazy Evaluation A x

At x.odd x.even

la_object_t foo(la_object_t A, la_object_t x) { x2

// At = A’
la_object_t At = la_transpose(A);
!

// sum odd elements of x to even elements of x

la_object_t x2 = la_sum(la_vector_slice(x,0,2,la_vector_length(x)/2),
la_vector_slice(x,1,2,la_vector_length(x)/2));
!

// Atx2 = A’ * x2 * 3.2
la_object_t Atx2 = la_scale_with_float(la_matrix_product(At,x2), 3.2f);
if (la_status(Atx2) < 0) { // error }
return Atx2
}
Lazy Evaluation A x

At x.odd x.even

la_object_t foo(la_object_t A, la_object_t x) { 3.2 x2

// At = A’
la_object_t At = la_transpose(A); Atx2
!

// sum odd elements of x to even elements of x

la_object_t x2 = la_sum(la_vector_slice(x,0,2,la_vector_length(x)/2),
la_vector_slice(x,1,2,la_vector_length(x)/2));
!

// Atx2 = A’ * x2 * 3.2
la_object_t Atx2 = la_scale_with_float(la_matrix_product(At,x2), 3.2f);
if (la_status(Atx2) < 0) { // error }
return Atx2
}
Lazy Evaluation
Details

No computation
No data buffer allocation
Triggered by
- la_matrix_to_float_buffer
- la_matrix_to_double_buffer
- la_vector_to_float_buffer
- la_vector_to_double_buffer
Performance Comparison

Netlib BLAS
• Open source
Performance Comparison
Higher is better
50 LA
Accelerate BLAS
Netlib BLAS

37.5
GFLOPS

25

12.5

32 160 288 416 544 672 800 928 1024

Matrix Size
Performance Comparison
Higher is better
50 LA
Accelerate BLAS
Netlib BLAS

37.5
GFLOPS

25

12.5

32 160 288 416 544 672 800 928 1024

Matrix Size
Performance Comparison
Higher is better
50 LA
Accelerate BLAS
Netlib BLAS

37.5
GFLOPS

25

12.5

32 160 288 416 544 672 800 928 1024

Matrix Size
Performance Comparison
Higher is better
50 LA
Accelerate BLAS
Netlib BLAS

37.5
GFLOPS

25

12.5

32 160 288 416 544 672 800 928 1024

Matrix Size
Error Handling

la_object_t AB = la_matrix_product( A, la_transpose(B) );

if (la_status(AB) < 0) { // handle error }
!

la_object_t result = la_sum( AB, la_scale_with_float( C, 3.2f ) );

if (la_status(result) < 0) { // handle error }
!

la_status_t status = la_matrix_to_float_buffer(buffer, leading_dim, result);

Error Handling

la_object_t AB = la_matrix_product( A, la_transpose(B) );

if (la_status(AB) < 0) { // handle error }
!

la_object_t result = la_sum( AB, la_scale_with_float( C, 3.2f ) );

if (la_status(result) < 0) { // handle error }
!

la_status_t status = la_matrix_to_float_buffer(buffer, leading_dim, result);

Error Handling

la_object_t AB = la_matrix_product( A, la_transpose(B) );

!

la_object_t result = la_sum( AB, la_scale_with_float( C, 3.2f ) );

!

la_status_t status = la_matrix_to_float_buffer(buffer, leading_dim, result);

if (status == LA_SUCCESS) {
// No errors, buffer is filled with good data.
} else if (status > 0) {
// No errors occurred, but result does not have full accuracy.
} else {
// An error occurred.
assert(0);
}
Debugging

Enable logging
• LA_ATTRIBUTE_ENABLE_LOGGING
Debugging

Enable logging
• LA_ATTRIBUTE_ENABLE_LOGGING

Error log
la_object_t la_sum(la_object_t, la_object_t):
LA_DIMENSION_MISMATCH_ERROR: Encountered a dimension mismatch
obj_left rows must be equal to obj_right rows; failed comparison: 8 == 9
Solve

la_object_t x = la_solve(A,b);
• If A is square and non-singular, compute the solution x to Ax = b
• If A is square and singular, produce an error
Slicing
What is slicing

Light weight access to partial object

• No buffer allocations
• No buffer copies
Slicing
What is slicing

Light weight access to partial object

• No buffer allocations
• No buffer copies
Three pieces of information
• Offset
• Stride
• Dimension
Slicing
What is slicing

la_vector_slice (vector, [0123456789]

7, // offset
-2, // stride
3); // dimension
Slicing
What is slicing

la_vector_slice (vector, [0123456789]

7, // offset
-2, // stride
3); // dimension [ 7 ]
Slicing
What is slicing

la_vector_slice (vector, [0123456789]

7, // offset
-2, // stride
3); // dimension [ 7 5 ]
Slicing
What is slicing

la_vector_slice (vector, [0123456789]

7, // offset
-2, // stride
3); // dimension [ 7 5 3 ]
Slice Example
Tiling

la_object_t A,B,C;
// A and B are matrices of dimension MxN
!

for (int i = 0; i < 2; ++i) {

for (int j = 0; j < 2; ++j) {
la_object_t Atile = la_matrix_slice(A,i*M/2,j*N/2,1,1,M/2,N/2);
la_object_t Btile = la_matrix_slice(B,i*M/2,j*N/2,1,1,M/2,N/2);
C = la_sum(Atile, Btile);
// use of C tile
}
}
Slice Example
Tiling

la_object_t A,B,C;
// A and B are matrices of dimension MxN
!

for (int i = 0; i < 2; ++i) {

for (int j = 0; j < 2; ++j) {
la_object_t Atile = la_matrix_slice(A,i*M/2,j*N/2,1,1,M/2,N/2);
la_object_t Btile = la_matrix_slice(B,i*M/2,j*N/2,1,1,M/2,N/2);
C = la_sum(Atile, Btile);
// use of C tile
}
}

C = +
A B
Slice Example
Tiling

la_object_t A,B,C;
// A and B are matrices of dimension MxN
!

for (int i = 0; i < 2; ++i) {

for (int j = 0; j < 2; ++j) {
la_object_t Atile = la_matrix_slice(A,i*M/2,j*N/2,1,1,M/2,N/2);
la_object_t Btile = la_matrix_slice(B,i*M/2,j*N/2,1,1,M/2,N/2);
C = la_sum(Atile, Btile);
// use of C tile
}
}

C = +
A B
Slice Example
Tiling

la_object_t A,B,C;
// A and B are matrices of dimension MxN
!

la_object_t sum = la_sum(A,B);

!

for (int i = 0; i < 2; ++i) {

for (int j = 0; j < 2; ++j) {
C = la_matrix_slice(sum,i*M/2,j*N/2,1,1,M/2,N/2);
// use of C tile
}
}
Slice Example
Tiling

la_object_t A,B,C;
// A and B are matrices of dimension MxN
!

la_object_t sum = la_sum(A,B);

!

for (int i = 0; i < 2; ++i) {

for (int j = 0; j < 2; ++j) {
C = la_matrix_slice(sum,i*M/2,j*N/2,1,1,M/2,N/2);
// use of C tile
}
}
Slice Example
Tiling

la_object_t A,B,C;
// A and B are matrices of dimension MxN
!

la_object_t sum = la_sum(A,B);

!

for (int i = 0; i < 2; ++i) {

for (int j = 0; j < 2; ++j) {
C = la_matrix_slice(sum,i*M/2,j*N/2,1,1,M/2,N/2);
// use of C tile
}
}

C =
sum
Slice Example
Tiling

la_object_t A,B,C;
// A and B are matrices of dimension MxN
!

la_object_t sum = la_sum(A,B);

!

for (int i = 0; i < 2; ++i) {

for (int j = 0; j < 2; ++j) {
C = la_matrix_slice(sum,i*M/2,j*N/2,1,1,M/2,N/2);
// use of C tile
}
}

C = = +
sum A B
Splat
What is splat

la_splat_from_float(5.0f );
Splat
What is splat

la_splat_from_float(5.0f );
Splat
What is splat

la_splat_from_float(5.0f );

Add 2 to every element of a vector

• la_sum(vector, la_splat_from_double(2.0));
LinearAlgebra
Summary

Simple API
Modern language and run-time features
Good performance
LINPACK
The joy of benchmarking

Stephen Canon
Engineer, Vector and Numerics Group
 How fast can you solve a 

system of linear equations?
 LINPACK tests both 

hardware and software
Accelerate vs. “Brand A”
 2013
LINPACK performance in GFLOPS (bigger is better)

Accelerate on iPhone 5

“Brand A”

1 2 3 4
 2014
LINPACK performance in GFLOPS (bigger is better)

Accelerate on iPhone 5

“Brand A”

1 2 3 4
Let’s find some new competition

Ourselves
Let’s find some new competition

Ourselves
2014
LINPACK performance in GFLOPS (bigger is better)

iPhone 5
10.4 GFLOPS

2010 MacBook Air

1 2 3 4 5 6 7 8
2014
LINPACK performance in GFLOPS (bigger is better)

iPhone 5s 10.4 GFLOPS

2010 MacBook Air

1 2 3 4 5 6 7 8 9 10 11
2014
LINPACK performance in GFLOPS (bigger is better)

iPhone 5s 10.4 GFLOPS

2010 MacBook Air

1 2 3 4 5 6 7 8 9 10 11 12 13
2014
LINPACK performance in GFLOPS (bigger is better)

iPad Air 14.6 GFLOPS

2010 MacBook Air

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
<simd/simd.h>
Short vector and matrix math
<simd/simd.h>

New (iOS 8 and OS X Yosemite) library with three purposes:

<simd/simd.h>

New (iOS 8 and OS X Yosemite) library with three purposes:

• 2D, 3D, and 4D vector math and geometry
<simd/simd.h>

New (iOS 8 and OS X Yosemite) library with three purposes:

• 2D, 3D, and 4D vector math and geometry
• Features of Metal in C, Objective-C, and C++ on the CPU
<simd/simd.h>

New (iOS 8 and OS X Yosemite) library with three purposes:

• 2D, 3D, and 4D vector math and geometry
• Features of Metal in C, Objective-C, and C++ on the CPU
• Abstraction over architecture-specific SIMD types and intrinsics
Vector Math and Geometry
Wish list
Vector Math and Geometry
Wish list

Inline implementations
Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
!

float a = cblas_sdot(3, 1.0f, x, 1, y, 1);

Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
!

float a = GLKVector3DotProduct(x, y);

Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
!

float a = vector_dot(x, y);

Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
!

using namespace simd;

float a = dot(x, y);
Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
Arithmetic should use operators
Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
Arithmetic should use operators
!

z = GLKVector4MultiplyScalar(GLKVector4Add(x,y),0.5);
Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
Arithmetic should use operators
!

z = 0.5*(x + y);
Vector Math and Geometry
Wish list

Inline implementations
Concise functions without extra parameters
Arithmetic should use operators
!

z = 0.5*(x + y);
Vector Math and Geometry
Types

In C and Objective-C, the primary type is vector_floatN, where N is 2, 3, or 4

In C++ you can use simd::floatN
Based on clang “extended vectors”
Vector Math and Geometry
Arithmetic

Your favorite arithmetic operators (+,–,*,/) work with both vectors and scalars
Vector Math and Geometry
Arithmetic

Your favorite arithmetic operators (+,–,*,/) work with both vectors and scalars
vector_float3 vector_reflect(vector_float3 x, vector_float3 n) {
!

}
Vector Math and Geometry
Arithmetic

Your favorite arithmetic operators (+,–,*,/) work with both vectors and scalars
vector_float3 vector_reflect(vector_float3 x, vector_float3 n) {
!

x
Vector Math and Geometry
Arithmetic

Your favorite arithmetic operators (+,–,*,/) work with both vectors and scalars
vector_float3 vector_reflect(vector_float3 x, vector_float3 n) {
!

x
n
Vector Math and Geometry
Arithmetic

Your favorite arithmetic operators (+,–,*,/) work with both vectors and scalars
vector_float3 vector_reflect(vector_float3 x, vector_float3 n) {
!

x
n
Vector Math and Geometry
Arithmetic

Your favorite arithmetic operators (+,–,*,/) work with both vectors and scalars
vector_float3 vector_reflect(vector_float3 x, vector_float3 n) {
!

x
n

vector_reflect(x,n)
Vector Math and Geometry
Arithmetic

Your favorite arithmetic operators (+,–,*,/) work with both vectors and scalars
vector_float3 vector_reflect(vector_float3 x, vector_float3 n) {
return x - 2*vector_dot(x,n)*n;
}

x
n

vector_reflect(x,n)
Vector Math and Geometry
Elements and subvectors

Array subscripting
vector_float4 a = { 0, 1, 2, 3 };
float z = x[2]; // z = 2
Vector Math and Geometry
Elements and subvectors

Named subvectors
vector_float4 a = { 0, 1, 2, 3 };
vector_float2 b = a.lo; // b = { 0, 1 }
a.even = -b; // a = { 0, 1,-1, 3 }
Vector Math and Geometry

<simd/math.h>

<simd/common.h>

<simd/geometry.h>
Vector Math and Geometry

C/Objective-C
fabs(x)
sqrt(x)
<simd/math.h>
floor(x)
sin(x)…
vector_clamp(x,min,max)
vector_mix(x,y,t)
<simd/common.h>
vector_recip(x)
vector_step(x,edge)…
vector_dot(x,y)
vector_length(x)
<simd/geometry.h>
vector_normalize(x)
vector_reflect(x,n)…
Vector Math and Geometry

C/Objective-C C++ (and Metal)

fabs(x) fabs(x)
sqrt(x) sqrt(x)
<simd/math.h>
floor(x) floor(x)
sin(x)… sin(x)…
vector_clamp(x,min,max) clamp(x,min,max)
vector_mix(x,y,t) mix(x,y,t)
<simd/common.h>
vector_recip(x) recip(x)
vector_step(x,edge)… step(x,edge)…
vector_dot(x,y) dot(x,y)
vector_length(x) length(x)
<simd/geometry.h>
vector_normalize(x) normalize(x)
vector_reflect(x,n)… reflect(x,n)…
Vector Math and Geometry

Some functions have two flavors: “precise” and “fast”

Vector Math and Geometry

Some functions have two flavors: “precise” and “fast”

• “precise” is the default…
Vector Math and Geometry

Some functions have two flavors: “precise” and “fast”

• “precise” is the default…
• … but if you compile with -ffast-math, you get the “fast” versions
Vector Math and Geometry

Even with -ffast-math, you can call the precise functions by name:
!

float len = vector_precise_length(x);

Vector Math and Geometry

You can call the fast versions by name too:

!

x = fast::normalize(x);
Matrices
C and Objective-C

“matrix_floatNxM”, where N and M are 2, 3, or 4

• N is number of columns, M is number of rows.
Matrices
C and Objective-C

Create matrices
matrix_from_diagonal(vector)
matrix_from_columns(vector, vector, …)
matrix_from_rows(vector, vector, …)
Arithmetic
matrix_scale(matrix, scalar)
matrix_linear_combination(matrix, scalar, matrix, scalar)
matrix_transpose(matrix)
matrix_invert(matrix)
matrix_multiply(matrix/vector, matrix/vector)
Matrices
C++ (and Metal)

Create matrices
float4x4()
float2x2(diagonal)
float3x4(column0, column1, …)
Arithmetic
float3x4 A, B;
A -= 2.f * B;
float4x3 C = transpose(A)
vector3 x;
vector4 y = A*x;
Abstract SIMD
Additional types

Doubles, signed and unsigned integers

Longer vectors (8, 16, and 32 elements)
Unaligned vectors
Abstract SIMD
Integer operators and conversions
Abstract SIMD
Integer operators and conversions

Arithmetic operators: +, -, *, /, %
Abstract SIMD
Integer operators and conversions

Arithmetic operators: +, -, *, /, %
Bitwise operators: >>, <<, &, |, ^, ~
Abstract SIMD
Integer operators and conversions

Arithmetic operators: +, -, *, /, %
Bitwise operators: >>, <<, &, |, ^, ~
Conversions:
vector_float x;
vector_ushort y = vector_ushort(x);
vector_char z = vector_char_sat(x);
Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false
Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false

x 0.0 1.0 2.0 3.0

Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false

x 0.0 1.0 2.0 3.0

y 0.0 3.14159 –infinity 42.0
Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false

x 0.0 1.0 2.0 3.0

y 0.0 3.14159 –infinity 42.0

x < y
Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false

x 0.0 1.0 2.0 3.0

y 0.0 3.14159 –infinity 42.0

x < y 0x00000000
Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false

x 0.0 1.0 2.0 3.0

y 0.0 3.14159 –infinity 42.0

x < y 0x00000000 0xffffffff

Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false

x 0.0 1.0 2.0 3.0

y 0.0 3.14159 –infinity 42.0

x < y 0x00000000 0xffffffff 0x00000000

Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false

x 0.0 1.0 2.0 3.0

y 0.0 3.14159 –infinity 42.0

x < y 0x00000000 0xffffffff 0x00000000 0xffffffff

Abstract SIMD
Comparisons

Vector comparisons: ==, !=, >, <, >=, <=

• Result is a vector of integers; each lane is –1 if comparison is true, 0 if false
• Type of result usually isn’t important, because you’ll use one of the following:
if (vector_any(x < 0)) { /* executed if any lane of x is negative */ }
if (vector_all(y != 0)) { /* executed if every lane of y is non-zero */ }
z = vector_bitselect(x, y, x > y); /* minimum of x and y */
String Copy
Scalar implementation
String Copy
Scalar implementation

void string_copy(char *dst, const char *src) {

while ((*dst++ = *src++));
}
String Copy
SSE intrinsic implementation

void vector_string_copy(char *dst, const char *src) {

while ((uintptr_t)src % 16)
if ((*dst++ = *src++) == 0) return;
while (1) {
__m128i data = _mm_load_si128((const __m128i *)src);
__m128i contains_zero = _mm_cmpeq_epi8(data, _mm_set1_epi8(0));
if (_mm_movemask_epi8(contains_zero))
break;
_mm_storeu_si128((__m128i *)dst, data);
src += 16;
dst += 16;
}
string_copy((char *)vec_dst, (const char *)vec_src);
}
String Copy
<simd/simd.h> implementation

void vector_string_copy(char *dst, const char *src) {

while ((uintptr_t)src % 16)
if ((*dst++ = *src++) == 0) return;
const vector_char16 *vec_src = (const vector_char16 *)src;
packed_char16 *vec_dst = (packed_char16 *)dst;
while (!vector_any(*vec_src == 0))
*vec_dst++ = *vec_src++;
string_copy((char *)vec_dst, (const char *)vec_src);
}
String Copy
<simd/simd.h> implementation

void vector_string_copy(char *dst, const char *src) {

while ((uintptr_t)src % 16)
if ((*dst++ = *src++) == 0) return;
const vector_char16 *vec_src = (const vector_char16 *)src;
packed_char16 *vec_dst = (packed_char16 *)dst;
while (!vector_any(*vec_src == 0))
*vec_dst++ = *vec_src++;
string_copy((char *)vec_dst, (const char *)vec_src);
}
String Copy
<simd/simd.h> implementation

void vector_string_copy(char *dst, const char *src) {

while ((uintptr_t)src % 16)
if ((*dst++ = *src++) == 0) return;
const vector_char16 *vec_src = (const vector_char16 *)src;
packed_char16 *vec_dst = (packed_char16 *)dst;
while (!vector_any(*vec_src == 0))
*vec_dst++ = *vec_src++;
string_copy((char *)vec_dst, (const char *)vec_src);
}
String Copy
<simd/simd.h> implementation

void vector_string_copy(char *dst, const char *src) {

while ((uintptr_t)src % 16)
if ((*dst++ = *src++) == 0) return;
const vector_char16 *vec_src = (const vector_char16 *)src;
packed_char16 *vec_dst = (packed_char16 *)dst;
while (!vector_any(*vec_src == 0))
*vec_dst++ = *vec_src++;
string_copy((char *)vec_dst, (const char *)vec_src);
}
String Copy
<simd/simd.h> implementation

void vector_string_copy(char *dst, const char *src) {

while ((uintptr_t)src % 16)
if ((*dst++ = *src++) == 0) return;
const vector_char16 *vec_src = (const vector_char16 *)src;
packed_char16 *vec_dst = (packed_char16 *)dst;
while (!vector_any(*vec_src == 0))
*vec_dst++ = *vec_src++;
string_copy((char *)vec_dst, (const char *)vec_src);
}
Performance
Higher is better
6 Scalar
<simd/simd.h>
Libc
Bytes copied per nanosecond

4.5

1.5

1 2 4 8 16 32 64 128 256 512 1024

String length in bytes
Performance
Higher is better
6 Scalar
<simd/simd.h>
Libc
Bytes copied per nanosecond

4.5

1.5

1 2 4 8 16 32 64 128 256 512 1024

String length in bytes
Performance
Higher is better
6 Scalar
<simd/simd.h>
Libc
Bytes copied per nanosecond

4.5

1.5

1 2 4 8 16 32 64 128 256 512 1024

String length in bytes
Performance
Higher is better
6 Scalar
<simd/simd.h>
Libc
Bytes copied per nanosecond

4.5

1.5

1 2 4 8 16 32 64 128 256 512 1024

String length in bytes
Summary

Simple interfaces for complex operations

New libraries still have a few rough edges
Let us know what use cases matter to you, and what additional features you need
More Information

Paul Danbold
Core OS Technology Evangelist
[email protected]
George Warner
DTS Sr. Support Scientist
[email protected]
Documentation
vImage Programming Guide
http://developer.apple.com/library/mac/#documentation/Performance/Conceptual/
vImage/Introduction/Introduction.html
More Information

Documentation
vDSP Programming Guide
http://developer.apple.com/library/mac/#documentation/Performance/Conceptual/
vDSP_Programming_Guide/Introduction/Introduction.html
vImage Headers
/System/Library/Frameworks/Accelerate.framework/Frameworks/vImage.framework/
Headers/vImage.h
vDSP Headers
/System/Library/Frameworks/Accelerate.framework/Frameworks/vecLib.framework/
Headers/vDSP.h
More Information

Documentation
LinearAlgebra Headers
/System/Library/Frameworks/Accelerate.framework/Frameworks/vecLib.framework/
Headers/LinearAlgebra/LinearAlgebra.h
<simd/simd.h>
/usr/include/simd/simd.h
Apple Developer Forums
http://devforums.apple.com
Bug Report
http://bugreport.apple.com
Related Sessions

• Working with Metal: Overview Pacific Heights Wednesday 9:00AM

• Working with Metal: Fundamentals Pacific Heights Wednesday 10:15AM

• Working with Metal: Advanced Pacific Heights Wednesday 11:30AM

Labs

• Accelerate Lab Core OS Lab B Tuesday 11:30AM

• Metal Lab Graphics and Games Lab A Wednesday 2:00PM