/* |
File: DemonstrateConvolution.c |
Abstract: Demonstration of vDSP convolution routine. |
Version: 1.2 |
|
Disclaimer: IMPORTANT: This Apple software is supplied to you by Apple |
Inc. ("Apple") in consideration of your agreement to the following |
terms, and your use, installation, modification or redistribution of |
this Apple software constitutes acceptance of these terms. If you do |
not agree with these terms, please do not use, install, modify or |
redistribute this Apple software. |
|
In consideration of your agreement to abide by the following terms, and |
subject to these terms, Apple grants you a personal, non-exclusive |
license, under Apple's copyrights in this original Apple software (the |
"Apple Software"), to use, reproduce, modify and redistribute the Apple |
Software, with or without modifications, in source and/or binary forms; |
provided that if you redistribute the Apple Software in its entirety and |
without modifications, you must retain this notice and the following |
text and disclaimers in all such redistributions of the Apple Software. |
Neither the name, trademarks, service marks or logos of Apple Inc. may |
be used to endorse or promote products derived from the Apple Software |
without specific prior written permission from Apple. Except as |
expressly stated in this notice, no other rights or licenses, express or |
implied, are granted by Apple herein, including but not limited to any |
patent rights that may be infringed by your derivative works or by other |
works in which the Apple Software may be incorporated. |
|
The Apple Software is provided by Apple on an "AS IS" basis. APPLE |
MAKES NO WARRANTIES, EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION |
THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY AND FITNESS |
FOR A PARTICULAR PURPOSE, REGARDING THE APPLE SOFTWARE OR ITS USE AND |
OPERATION ALONE OR IN COMBINATION WITH YOUR PRODUCTS. |
|
IN NO EVENT SHALL APPLE BE LIABLE FOR ANY SPECIAL, INDIRECT, INCIDENTAL |
OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF |
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS |
INTERRUPTION) ARISING IN ANY WAY OUT OF THE USE, REPRODUCTION, |
MODIFICATION AND/OR DISTRIBUTION OF THE APPLE SOFTWARE, HOWEVER CAUSED |
AND WHETHER UNDER THEORY OF CONTRACT, TORT (INCLUDING NEGLIGENCE), |
STRICT LIABILITY OR OTHERWISE, EVEN IF APPLE HAS BEEN ADVISED OF THE |
POSSIBILITY OF SUCH DAMAGE. |
|
Copyright (C) 2012 Apple Inc. All Rights Reserved. |
|
|
This is a sample module to illustrate the use of vDSP_conv for |
convolution and correlation. This module also times the vDSP_conv. |
|
Copyright (C) 2007 Apple Inc. All rights reserved. |
*/ |
|
#include <stdio.h> |
#include <stdlib.h> |
|
#include <Accelerate/Accelerate.h> |
|
#include "Demonstrate.h" |
|
|
#define Iterations 1000 // How many iterations to use in the timer loop. |
|
|
// Demonstrate vDSP_conv. |
void DemonstrateConvolution(void) |
{ |
/* Define strides for the arrays being passed to the convolution. |
In many applications, the strides are one and are passed to the |
vDSP routine as constants. We give them names here just so |
that the routine call is more illustrative, showing the |
arguments by name. |
*/ |
const vDSP_Stride |
SignalStride = 1, |
FilterStride = 1, |
ResultStride = 1; |
|
/* Lengths tend to differ more than strides, although it is might |
not be unusual to have constants in a particular application. |
(For example, your filter length might be a constant because |
the code is written specially for a very specific filter. The |
result length is more likely to be a named constant at least, |
representing a frame size.) |
|
The signal length is padded a bit. This length is not actually |
passed to the vDSP_conv routine; it is the number of elements |
that the signal array must contain. The SignalLength defined |
below is used to allocate space, and it is the filter length |
rounded up to a multiple of four elements and added to the |
result length. The extra elements give the vDSP_conv routine |
leeway to perform vector-load instructions, which load multiple |
elements even if they are not all used. If the caller did not |
guarantee that memory beyond the values used in the signal |
array were accessible, a memory access violation might result. |
*/ |
vDSP_Length |
FilterLength = 256, |
ResultLength = 2048, |
SignalLength = (FilterLength+3 & -4u) + ResultLength; |
|
// Define pointers to the arrays we will use. |
float *Signal, *Filter, *Result; |
|
// Define a variable for a loop iterator. |
vDSP_Length i; |
|
// Define some variables used to time the routine. |
ClockData t0, t1; |
float Time, Gigaflops; |
|
printf("Begin %s.\n\n", __func__); |
|
// Allocate memory for the arrays. |
Signal = malloc(SignalLength * SignalStride * sizeof *Signal); |
Filter = malloc(FilterLength * FilterStride * sizeof *Filter); |
Result = malloc(ResultLength * ResultStride * sizeof *Result); |
|
if (Signal == NULL || Filter == NULL || Result == NULL) |
{ |
fprintf(stderr, "Error, failed to allocate memory.\n"); |
exit(EXIT_FAILURE); |
} |
|
/* Initialize the input signal. In a real application, data would |
of course be provided from an image file, sensors, or other |
source. |
*/ |
for (i = 0; i < SignalLength; i++) |
Signal[i*SignalStride] = 1.0; |
|
/* Initialize the filter. A real application would have different |
filter coefficients. |
*/ |
for (i = 0; i < FilterLength; i++) |
Filter[i*FilterStride] = 1.0; |
|
// Perform a correlation. |
vDSP_conv(Signal, SignalStride, Filter, FilterStride, |
Result, ResultStride, ResultLength, FilterLength); |
|
/* Perform a convolution by using the filter backward. To do |
this, we pass -1 for the filter stride and pass vDSP_conv a |
pointer to the first element of the backward filter, which is |
the last element of the forward filter. |
*/ |
vDSP_conv(Signal, SignalStride, Filter + FilterLength - 1, -1, |
Result, ResultStride, ResultLength, FilterLength); |
|
/* The above calls show how to use the vDSP_conv routine. Now we |
will see how fast it is. |
*/ |
|
t0 = Clock(); |
|
for (i = 0; i < Iterations; i++) |
vDSP_conv(Signal, SignalStride, Filter, FilterStride, |
Result, ResultStride, ResultLength, FilterLength); |
|
t1 = Clock(); |
|
// Average the time over all the loop iterations. |
Time = ClockToSeconds(t1, t0) / Iterations; |
|
/* For each result element, a convolution takes a multiply for |
each filter element and an addition for each element after the |
first. So there are ResultLength * (2 * FilterLength - 1) |
floating-point operations in this convolution. We divide that |
by the number of seconds it takes to do the convolution to get |
floating-point operations per second and then scale by 1e-9 to |
get gigaflops. |
*/ |
Gigaflops = ResultLength * (2 * FilterLength - 1) / Time * 1e-9; |
|
printf("\tA %lu * %lu convolution takes %g microseconds,\n" |
"\twhich is a performance of %g gigaflops.\n\n", |
(unsigned long) ResultLength, (unsigned long) FilterLength, |
Time * 1e6, Gigaflops); |
|
// Free allocated memory. |
free(Signal); |
free(Filter); |
free(Result); |
|
printf("End %s.\n\n\n", __func__); |
} |
