Blame - docs/00_introduction.dox - ml/ComputeLibrary

blob: 50efb518af0337fa850f8cd0305c5a309b1dcf4d [file] [log] [blame]

Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	1	/** @mainpage Introduction
				2
				3	@tableofcontents
				4
				5	The Computer Vision and Machine Learning library is a set of functions optimised for both ARM CPUs and GPUs using SIMD technologies.
				6
				7	Several builds of the library are available using various configurations:
				8	- OS: Linux, Android or bare metal.
				9	- Architecture: armv7a (32bit) or arm64-v8a (64bit)
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	10	- Technology: NEON / OpenCL / GLES_COMPUTE / NEON and OpenCL and GLES_COMPUTE
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	11	- Debug / Asserts / Release: Use a build with asserts enabled to debug your application and enable extra validation. Once you are sure your application works as expected you can switch to a release build of the library for maximum performance.
				12
				13	@section S0_1_contact Contact / Support
				14
				15	Please email developer@arm.com
				16
				17	In order to facilitate the work of the support team please provide the build information of the library you are using. To get the version of the library you are using simply run:
				18
				19	$ strings android-armv7a-cl-asserts/libarm_compute.so \| grep arm_compute_version
				20	arm_compute_version=v16.12 Build options: {'embed_kernels': '1', 'opencl': '1', 'arch': 'armv7a', 'neon': '0', 'asserts': '1', 'debug': '0', 'os': 'android', 'Werror': '1'} Git hash=f51a545d4ea12a9059fe4e598a092f1fd06dc858
				21
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	22	@section S0_2_prebuilt_binaries Pre-built binaries
				23
				24	For each release we provide some pre-built binaries of the library [here](https://github.com/ARM-software/ComputeLibrary/releases)
				25
				26	These binaries have been built using the following toolchains:
				27	- Linux armv7a: gcc-linaro-arm-linux-gnueabihf-4.9-2014.07_linux
				28	- Linux arm64-v8a: gcc-linaro-4.9-2016.02-x86_64_aarch64-linux-gnu
Anthony Barbier	a8a28f6	2018-02-26 19:16:32 +0000	[diff] [blame]	29	- Android armv7a: clang++ / gnustl NDK r16b
				30	- Android am64-v8a: clang++ / gnustl NDK r16b
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	31
				32	@warning Make sure to use a compatible toolchain to build your application or you will get some std::bad_alloc errors at runtime.
				33
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	34	@section S1_file_organisation File organisation
				35
				36	This archive contains:
				37	- The arm_compute header and source files
				38	- The latest Khronos OpenCL 1.2 C headers from the <a href="https://www.khronos.org/registry/cl/">Khronos OpenCL registry</a>
				39	- The latest Khronos cl2.hpp from the <a href="https://www.khronos.org/registry/cl/">Khronos OpenCL registry</a> (API version 2.1 when this document was written)
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	40	- The latest Khronos OpenGL ES 3.1 C headers from the <a href="https://www.khronos.org/registry/gles/">Khronos OpenGL ES registry</a>
				41	- The latest Khronos EGL 1.5 C headers from the <a href="https://www.khronos.org/registry/gles/">Khronos EGL registry</a>
				42	- The sources for a stub version of libOpenCL.so, libGLESv1_CM.so, libGLESv2.so and libEGL.so to help you build your application.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	43	- An examples folder containing a few examples to compile and link against the library.
				44	- A @ref utils folder containing headers with some boiler plate code used by the examples.
				45	- This documentation.
				46
				47	You should have the following file organisation:
				48
				49	.
				50	├── arm_compute --> All the arm_compute headers
				51	│ ├── core
				52	│ │ ├── CL
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	53	│ │ │ ├── CLKernelLibrary.h --> Manages all the OpenCL kernels compilation and caching, provides accessors for the OpenCL Context.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	54	│ │ │ ├── CLKernels.h --> Includes all the OpenCL kernels at once
				55	│ │ │ ├── CL specialisation of all the generic objects interfaces (ICLTensor, ICLImage, etc.)
				56	│ │ │ ├── kernels --> Folder containing all the OpenCL kernels
				57	│ │ │ │ └── CL*Kernel.h
				58	│ │ │ └── OpenCL.h --> Wrapper to configure the Khronos OpenCL C++ header
				59	│ │ ├── CPP
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	60	│ │ │ ├── CPPKernels.h --> Includes all the CPP kernels at once
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	61	│ │ │ └── kernels --> Folder containing all the CPP kernels
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	62	│ │ │ └── CPP*Kernel.h
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	63	│ │ ├── GLES_COMPUTE
				64	│ │ │ ├── GCKernelLibrary.h --> Manages all the GLES kernels compilation and caching, provides accessors for the GLES Context.
				65	│ │ │ ├── GCKernels.h --> Includes all the GLES kernels at once
				66	│ │ │ ├── GLES specialisation of all the generic objects interfaces (IGCTensor, IGCImage, etc.)
				67	│ │ │ ├── kernels --> Folder containing all the GLES kernels
				68	│ │ │ │ └── GC*Kernel.h
				69	│ │ │ └── OpenGLES.h --> Wrapper to configure the Khronos EGL and OpenGL ES C header
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	70	│ │ ├── NEON
				71	│ │ │ ├── kernels --> Folder containing all the NEON kernels
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	72	│ │ │ │ ├── arm64 --> Folder containing the interfaces for the assembly arm64 NEON kernels
				73	│ │ │ │ ├── arm32 --> Folder containing the interfaces for the assembly arm32 NEON kernels
				74	│ │ │ │ ├── assembly --> Folder containing the NEON assembly routines.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	75	│ │ │ │ └── NE*Kernel.h
				76	│ │ │ └── NEKernels.h --> Includes all the NEON kernels at once
				77	│ │ ├── All common basic types (Types.h, Window, Coordinates, Iterator, etc.)
				78	│ │ ├── All generic objects interfaces (ITensor, IImage, etc.)
				79	│ │ └── Objects metadata classes (ImageInfo, TensorInfo, MultiImageInfo)
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	80	│ ├── graph
				81	│ │ ├── CL --> OpenCL specific operations
				82	│ │ │ └── CLMap.h / CLUnmap.h
				83	│ │ ├── nodes
				84	│ │ │ └── The various nodes supported by the graph API
				85	│ │ ├── Nodes.h --> Includes all the Graph nodes at once.
				86	│ │ └── Graph objects ( INode, ITensorAccessor, Graph, etc.)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	87	│ └── runtime
				88	│ ├── CL
				89	│ │ ├── CL objects & allocators (CLArray, CLImage, CLTensor, etc.)
				90	│ │ ├── functions --> Folder containing all the OpenCL functions
				91	│ │ │ └── CL*.h
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	92	│ │ ├── CLScheduler.h --> Interface to enqueue OpenCL kernels and get/set the OpenCL CommandQueue and ICLTuner.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	93	│ │ └── CLFunctions.h --> Includes all the OpenCL functions at once
				94	│ ├── CPP
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	95	│ │ ├── CPPKernels.h --> Includes all the CPP functions at once.
				96	│ │ └── CPPScheduler.h --> Basic pool of threads to execute CPP/NEON code on several cores in parallel
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	97	│ ├── GLES_COMPUTE
				98	│ │ ├── GLES objects & allocators (GCArray, GCImage, GCTensor, etc.)
				99	│ │ ├── functions --> Folder containing all the GLES functions
				100	│ │ │ └── GC*.h
				101	│ │ ├── GCScheduler.h --> Interface to enqueue GLES kernels and get/set the GLES CommandQueue.
				102	│ │ └── GCFunctions.h --> Includes all the GLES functions at once
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	103	│ ├── NEON
				104	│ │ ├── functions --> Folder containing all the NEON functions
				105	│ │ │ └── NE*.h
				106	│ │ └── NEFunctions.h --> Includes all the NEON functions at once
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	107	│ ├── OMP
				108	│ │ └── OMPScheduler.h --> OpenMP scheduler (Alternative to the CPPScheduler)
				109	│ ├── Memory manager files (LifetimeManager, PoolManager, etc.)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	110	│ └── Basic implementations of the generic object interfaces (Array, Image, Tensor, etc.)
Anthony Barbier	a8a28f6	2018-02-26 19:16:32 +0000	[diff] [blame]	111	├── data -> Contains test images and reference data dumps used by validation tests
				112	├── docs -> Contains Doxyfile and Doxygen sources used to generate the HTML pages in the documentation folder.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	113	├── documentation
				114	│ ├── index.xhtml
				115	│ └── ...
				116	├── documentation.xhtml -> documentation/index.xhtml
				117	├── examples
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	118	│ ├── cl_*.cpp --> OpenCL examples
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	119	│ ├── gc_*.cpp --> GLES compute shaders examples
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	120	│ ├── graph_*.cpp --> Graph examples
				121	│ ├── neoncl_*.cpp --> NEON / OpenCL interoperability examples
				122	│ └── neon_*.cpp --> NEON examples
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	123	├── include
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	124	│ ├── CL
				125	│ │ └── Khronos OpenCL C headers and C++ wrapper
				126	│ ├── half --> FP16 library available from http://half.sourceforge.net
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	127	│ ├── libnpy --> Library to load / write npy buffers, available from https://github.com/llohse/libnpy
				128	│ └── linux --> Headers only needed for Linux builds
				129	│ └── Khronos EGL and OpenGLES headers
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	130	├── opencl-1.2-stubs
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	131	│ └── opencl_stubs.c --> OpenCL stubs implementation
				132	├── opengles-3.1-stubs
				133	│ ├── EGL.c --> EGL stubs implementation
				134	│ └── GLESv2.c --> GLESv2 stubs implementation
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	135	├── scripts
				136	│ ├── caffe_data_extractor.py --> Basic script to export weights from Caffe to npy files
				137	│ └── tensorflow_data_extractor.py --> Basic script to export weights from Tensor Flow to npy files
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	138	├── src
				139	│ ├── core
				140	│ │ └── ... (Same structure as headers)
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	141	│ │ ├── CL
				142	│ │ │ └── cl_kernels --> All the OpenCL kernels
				143	│ │ └── GLES_COMPUTE
				144	│ │ └── cs_shaders --> All the OpenGL ES Compute Shaders
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	145	│ ├── graph
				146	│ │ └── ... (Same structure as headers)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	147	│ └── runtime
				148	│ └── ... (Same structure as headers)
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	149	├── support
				150	│ └── Various headers to work around toolchains / platform issues.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	151	├── tests
				152	│ ├── All test related files shared between validation and benchmark
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	153	│ ├── CL --> OpenCL accessors
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	154	│ ├── GLES_COMPUTE --> GLES accessors
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	155	│ ├── NEON --> NEON accessors
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	156	│ ├── benchmark --> Sources for benchmarking
				157	│ │ ├── Benchmark specific files
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	158	│ │ ├── CL --> OpenCL benchmarking tests
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	159	│ │ ├── GLES_COMPUTE --> GLES benchmarking tests
Anthony Barbier	a8a28f6	2018-02-26 19:16:32 +0000	[diff] [blame]	160	│ │ ├── fixtures
				161	│ │ │ └── Fixtures to initialise and run the runtime Functions.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	162	│ │ └── NEON --> NEON benchmarking tests
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	163	│ ├── datasets
				164	│ │ └── Datasets for all the validation / benchmark tests, layer configurations for various networks, etc.
				165	│ ├── framework
				166	│ │ └── Boiler plate code for both validation and benchmark test suites (Command line parsers, instruments, output loggers, etc.)
				167	│ ├── networks
				168	│ │ └── Examples of how to instantiate networks.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	169	│ ├── validation --> Sources for validation
				170	│ │ ├── Validation specific files
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	171	│ │ ├── CL --> OpenCL validation tests
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	172	│ │ ├── GLES_COMPUTE --> GLES validation tests
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	173	│ │ ├── CPP --> C++ reference implementations
				174	│ │ ├── fixtures
				175	│ │ │ └── Fixtures to initialise and run the runtime Functions.
				176	│ │ └── NEON --> NEON validation tests
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	177	│ └── dataset --> Datasets defining common sets of input parameters
				178	└── utils --> Boiler plate code used by examples
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	179	└── Various utilities to print types, load / store assets, etc.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	180
				181	@section S2_versions_changelog Release versions and changelog
				182
				183	@subsection S2_1_versions Release versions
				184
				185	All releases are numbered vYY.MM Where YY are the last two digits of the year, and MM the month number.
				186	If there is more than one release in a month then an extra sequential number is appended at the end:
				187
				188	v17.03 (First release of March 2017)
				189	v17.03.1 (Second release of March 2017)
				190	v17.04 (First release of April 2017)
				191
				192	@note We're aiming at releasing one major public release with new features per quarter. All releases in between will only contain bug fixes.
				193
				194	@subsection S2_2_changelog Changelog
				195
Anthony Barbier	2d0ce77	2018-02-21 15:35:36 +0000	[diff] [blame]	196	v18.02 Public major release
				197	- Various NEON / OpenCL / GLES optimisations.
				198	- Various bug fixes.
				199	- Changed default number of threads on big LITTLE systems.
				200	- Refactored examples and added:
				201	- graph_mobilenet_qassym8
				202	- graph_resnet
				203	- graph_squeezenet_v1_1
				204	- Renamed @ref arm_compute::CLConvolutionLayer into @ref arm_compute::CLGEMMConvolutionLayer and created a new @ref arm_compute::CLConvolutionLayer to select the fastest convolution method.
				205	- Renamed @ref arm_compute::NEConvolutionLayer into @ref arm_compute::NEGEMMConvolutionLayer and created a new @ref arm_compute::NEConvolutionLayer to select the fastest convolution method.
				206	- Added in place support to:
				207	- @ref arm_compute::CLActivationLayer
				208	- @ref arm_compute::CLBatchNormalizationLayer
				209	- Added QASYMM8 support to:
				210	- @ref arm_compute::CLActivationLayer
				211	- @ref arm_compute::CLDepthwiseConvolutionLayer
				212	- @ref arm_compute::NEDepthwiseConvolutionLayer
				213	- @ref arm_compute::NESoftmaxLayer
				214	- Added FP16 support to:
				215	- @ref arm_compute::CLDepthwiseConvolutionLayer3x3
				216	- @ref arm_compute::CLDepthwiseConvolutionLayer
				217	- Added broadcasting support to @ref arm_compute::NEArithmeticAddition / @ref arm_compute::CLArithmeticAddition / @ref arm_compute::CLPixelWiseMultiplication
				218	- Added fused batched normalization and activation to @ref arm_compute::CLBatchNormalizationLayer and @ref arm_compute::NEBatchNormalizationLayer
				219	- Added support for non-square pooling to @ref arm_compute::NEPoolingLayer and @ref arm_compute::CLPoolingLayer
				220	- New OpenCL kernels / functions:
				221	- @ref arm_compute::CLDirectConvolutionLayerOutputStageKernel
Pablo Tello	f6c572c	2018-02-14 12:47:30 +0000	[diff] [blame]	222	- New NEON kernels / functions
Anthony Barbier	2d0ce77	2018-02-21 15:35:36 +0000	[diff] [blame]	223	- Added name() method to all kernels.
				224	- Added support for Winograd 5x5.
				225	- @ref arm_compute::NEPermuteKernel / @ref arm_compute::NEPermute
Pablo Tello	f6c572c	2018-02-14 12:47:30 +0000	[diff] [blame]	226	- @ref arm_compute::NEWinogradLayerTransformInputKernel / @ref arm_compute::NEWinogradLayer
				227	- @ref arm_compute::NEWinogradLayerTransformOutputKernel / @ref arm_compute::NEWinogradLayer
				228	- @ref arm_compute::NEWinogradLayerTransformWeightsKernel / @ref arm_compute::NEWinogradLayer
				229	- Renamed arm_compute::NEWinogradLayerKernel into @ref arm_compute::NEWinogradLayerBatchedGEMMKernel
Anthony Barbier	2d0ce77	2018-02-21 15:35:36 +0000	[diff] [blame]	230	- New GLES kernels / functions:
				231	- @ref arm_compute::GCTensorShiftKernel / @ref arm_compute::GCTensorShift
Pablo Tello	f6c572c	2018-02-14 12:47:30 +0000	[diff] [blame]	232
Anthony Barbier	64c95a0	2018-01-22 18:48:55 +0000	[diff] [blame]	233	v18.01 Public maintenance release
				234	- Various bug fixes
				235	- Added some of the missing validate() methods
				236	- Added @ref arm_compute::CLDeconvolutionLayerUpsampleKernel / @ref arm_compute::CLDeconvolutionLayer @ref arm_compute::CLDeconvolutionLayerUpsample
				237	- Added @ref arm_compute::CLPermuteKernel / @ref arm_compute::CLPermute
				238	- Added method to clean the programs cache in the CL Kernel library.
				239	- Added @ref arm_compute::GCArithmeticAdditionKernel / @ref arm_compute::GCArithmeticAddition
				240	- Added @ref arm_compute::GCDepthwiseConvolutionLayer3x3Kernel / @ref arm_compute::GCDepthwiseConvolutionLayer3x3
				241	- Added @ref arm_compute::GCNormalizePlanarYUVLayerKernel / @ref arm_compute::GCNormalizePlanarYUVLayer
				242	- Added @ref arm_compute::GCScaleKernel / @ref arm_compute::GCScale
				243	- Added @ref arm_compute::GCWeightsReshapeKernel / @ref arm_compute::GCConvolutionLayer
				244	- Added FP16 support to the following GLES compute kernels:
				245	- @ref arm_compute::GCCol2ImKernel
				246	- @ref arm_compute::GCGEMMInterleave4x4Kernel
				247	- @ref arm_compute::GCGEMMTranspose1xWKernel
				248	- @ref arm_compute::GCIm2ColKernel
Pablo Tello	f6c572c	2018-02-14 12:47:30 +0000	[diff] [blame]	249	- Refactored NEON Winograd (arm_compute::NEWinogradLayerKernel)
Anthony Barbier	64c95a0	2018-01-22 18:48:55 +0000	[diff] [blame]	250	- Added @ref arm_compute::NEDirectConvolutionLayerOutputStageKernel
				251	- Added QASYMM8 support to the following NEON kernels:
				252	- @ref arm_compute::NEDepthwiseConvolutionLayer3x3Kernel
				253	- @ref arm_compute::NEFillBorderKernel
				254	- @ref arm_compute::NEPoolingLayerKernel
				255	- Added new examples:
				256	- graph_cl_mobilenet_qasymm8.cpp
				257	- graph_inception_v3.cpp
				258	- gc_dc.cpp
				259	- More tests added to both validation and benchmarking suites.
				260
Gian Marco	ff85093	2017-12-11 12:37:17 +0000	[diff] [blame]	261	v17.12 Public major release
				262	- Most machine learning functions on OpenCL support the new data type QASYMM8
				263	- Introduced logging interface
				264	- Introduced opencl timer
				265	- Reworked GEMMLowp interface
				266	- Added new NEON assembly kernels for GEMMLowp, SGEMM and HGEMM
				267	- Added validation method for most Machine Learning kernels / functions
				268	- Added new graph examples such as googlenet, mobilenet, squeezenet, vgg16 and vgg19
				269	- Added sgemm example for OpenCL
				270	- Added absolute difference example for GLES compute
				271	- Added new tests and benchmarks in validation and benchmark frameworks
				272	- Added new kernels / functions for GLES compute
				273
				274	- New OpenGL ES kernels / functions
				275	- @ref arm_compute::GCAbsoluteDifferenceKernel / @ref arm_compute::GCAbsoluteDifference
				276	- @ref arm_compute::GCActivationLayerKernel / @ref arm_compute::GCActivationLayer
				277	- @ref arm_compute::GCBatchNormalizationLayerKernel / @ref arm_compute::GCBatchNormalizationLayer
				278	- @ref arm_compute::GCCol2ImKernel
				279	- @ref arm_compute::GCDepthConcatenateLayerKernel / @ref arm_compute::GCDepthConcatenateLayer
				280	- @ref arm_compute::GCDirectConvolutionLayerKernel / @ref arm_compute::GCDirectConvolutionLayer
				281	- @ref arm_compute::GCDropoutLayerKernel / @ref arm_compute::GCDropoutLayer
				282	- @ref arm_compute::GCFillBorderKernel / @ref arm_compute::GCFillBorder
				283	- @ref arm_compute::GCGEMMInterleave4x4Kernel / @ref arm_compute::GCGEMMInterleave4x4
				284	- @ref arm_compute::GCGEMMMatrixAccumulateBiasesKernel / @ref arm_compute::GCGEMMMatrixAdditionKernel / @ref arm_compute::GCGEMMMatrixMultiplyKernel / @ref arm_compute::GCGEMM
				285	- @ref arm_compute::GCGEMMTranspose1xWKernel / @ref arm_compute::GCGEMMTranspose1xW
				286	- @ref arm_compute::GCIm2ColKernel
				287	- @ref arm_compute::GCNormalizationLayerKernel / @ref arm_compute::GCNormalizationLayer
				288	- @ref arm_compute::GCPixelWiseMultiplicationKernel / @ref arm_compute::GCPixelWiseMultiplication
				289	- @ref arm_compute::GCPoolingLayerKernel / @ref arm_compute::GCPoolingLayer
				290	- @ref arm_compute::GCLogits1DMaxKernel / @ref arm_compute::GCLogits1DShiftExpSumKernel / @ref arm_compute::GCLogits1DNormKernel / @ref arm_compute::GCSoftmaxLayer
				291	- @ref arm_compute::GCTransposeKernel / @ref arm_compute::GCTranspose
				292
				293	- New NEON kernels / functions
				294	- @ref arm_compute::NEGEMMLowpAArch64A53Kernel / @ref arm_compute::NEGEMMLowpAArch64Kernel / @ref arm_compute::NEGEMMLowpAArch64V8P4Kernel / arm_compute::NEGEMMInterleavedBlockedKernel / @ref arm_compute::NEGEMMLowpAssemblyMatrixMultiplyCore
				295	- @ref arm_compute::NEHGEMMAArch64FP16Kernel
				296	- @ref arm_compute::NEDepthwiseConvolutionLayer3x3Kernel / @ref arm_compute::NEDepthwiseIm2ColKernel / @ref arm_compute::NEGEMMMatrixVectorMultiplyKernel / @ref arm_compute::NEDepthwiseVectorToTensorKernel / @ref arm_compute::NEDepthwiseConvolutionLayer
				297	- @ref arm_compute::NEGEMMLowpOffsetContributionKernel / @ref arm_compute::NEGEMMLowpMatrixAReductionKernel / @ref arm_compute::NEGEMMLowpMatrixBReductionKernel / @ref arm_compute::NEGEMMLowpMatrixMultiplyCore
				298	- @ref arm_compute::NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel / @ref arm_compute::NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint
				299	- @ref arm_compute::NEGEMMLowpQuantizeDownInt32ToUint8ScaleKernel / @ref arm_compute::NEGEMMLowpQuantizeDownInt32ToUint8Scale
Pablo Tello	f6c572c	2018-02-14 12:47:30 +0000	[diff] [blame]	300	- @ref arm_compute::NEWinogradLayer / arm_compute::NEWinogradLayerKernel
Gian Marco	ff85093	2017-12-11 12:37:17 +0000	[diff] [blame]	301
				302	- New OpenCL kernels / functions
				303	- @ref arm_compute::CLGEMMLowpOffsetContributionKernel / @ref arm_compute::CLGEMMLowpMatrixAReductionKernel / @ref arm_compute::CLGEMMLowpMatrixBReductionKernel / @ref arm_compute::CLGEMMLowpMatrixMultiplyCore
				304	- @ref arm_compute::CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel / @ref arm_compute::CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint
				305	- @ref arm_compute::CLGEMMLowpQuantizeDownInt32ToUint8ScaleKernel / @ref arm_compute::CLGEMMLowpQuantizeDownInt32ToUint8Scale
				306
				307	- New graph nodes for NEON and OpenCL
				308	- @ref arm_compute::graph::BranchLayer
				309	- @ref arm_compute::graph::DepthConvertLayer
				310	- @ref arm_compute::graph::DepthwiseConvolutionLayer
				311	- @ref arm_compute::graph::DequantizationLayer
				312	- @ref arm_compute::graph::FlattenLayer
				313	- @ref arm_compute::graph::QuantizationLayer
				314	- @ref arm_compute::graph::ReshapeLayer
				315
Anthony Barbier	3c5b4ff	2017-10-12 13:20:52 +0100	[diff] [blame]	316	v17.10 Public maintenance release
				317	- Bug fixes:
				318	- Check the maximum local workgroup size supported by OpenCL devices
				319	- Minor documentation updates (Fixed instructions to build the examples)
				320	- Introduced a arm_compute::graph::GraphContext
				321	- Added a few new Graph nodes, support for branches and grouping.
				322	- Automatically enable cl_printf in debug builds
				323	- Fixed bare metal builds for armv7a
				324	- Added AlexNet and cartoon effect examples
				325	- Fixed library builds: libraries are no longer built as supersets of each other.(It means application using the Runtime part of the library now need to link against both libarm_compute_core and libarm_compute)
				326
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	327	v17.09 Public major release
				328	- Experimental Graph support: initial implementation of a simple stream API to easily chain machine learning layers.
				329	- Memory Manager (@ref arm_compute::BlobLifetimeManager, @ref arm_compute::BlobMemoryPool, @ref arm_compute::ILifetimeManager, @ref arm_compute::IMemoryGroup, @ref arm_compute::IMemoryManager, @ref arm_compute::IMemoryPool, @ref arm_compute::IPoolManager, @ref arm_compute::MemoryManagerOnDemand, @ref arm_compute::PoolManager)
				330	- New validation and benchmark frameworks (Boost and Google frameworks replaced by homemade framework).
				331	- Most machine learning functions support both fixed point 8 and 16 bit (QS8, QS16) for both NEON and OpenCL.
				332	- New NEON kernels / functions:
				333	- @ref arm_compute::NEGEMMAssemblyBaseKernel @ref arm_compute::NEGEMMAArch64Kernel
				334	- @ref arm_compute::NEDequantizationLayerKernel / @ref arm_compute::NEDequantizationLayer
				335	- @ref arm_compute::NEFloorKernel / @ref arm_compute::NEFloor
Giorgio Arena	04a8f8c	2017-11-23 11:45:24 +0000	[diff] [blame]	336	- @ref arm_compute::NEL2NormalizeLayerKernel / @ref arm_compute::NEL2NormalizeLayer
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	337	- @ref arm_compute::NEQuantizationLayerKernel @ref arm_compute::NEMinMaxLayerKernel / @ref arm_compute::NEQuantizationLayer
				338	- @ref arm_compute::NEROIPoolingLayerKernel / @ref arm_compute::NEROIPoolingLayer
				339	- @ref arm_compute::NEReductionOperationKernel / @ref arm_compute::NEReductionOperation
				340	- @ref arm_compute::NEReshapeLayerKernel / @ref arm_compute::NEReshapeLayer
				341
				342	- New OpenCL kernels / functions:
Giorgio Arena	04a8f8c	2017-11-23 11:45:24 +0000	[diff] [blame]	343	- @ref arm_compute::CLDepthwiseConvolutionLayer3x3Kernel @ref arm_compute::CLDepthwiseIm2ColKernel @ref arm_compute::CLDepthwiseVectorToTensorKernel @ref arm_compute::CLDepthwiseWeightsReshapeKernel / @ref arm_compute::CLDepthwiseConvolutionLayer3x3 @ref arm_compute::CLDepthwiseConvolutionLayer @ref arm_compute::CLDepthwiseSeparableConvolutionLayer
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	344	- @ref arm_compute::CLDequantizationLayerKernel / @ref arm_compute::CLDequantizationLayer
				345	- @ref arm_compute::CLDirectConvolutionLayerKernel / @ref arm_compute::CLDirectConvolutionLayer
				346	- @ref arm_compute::CLFlattenLayer
				347	- @ref arm_compute::CLFloorKernel / @ref arm_compute::CLFloor
				348	- @ref arm_compute::CLGEMMTranspose1xW
				349	- @ref arm_compute::CLGEMMMatrixVectorMultiplyKernel
Giorgio Arena	04a8f8c	2017-11-23 11:45:24 +0000	[diff] [blame]	350	- @ref arm_compute::CLL2NormalizeLayerKernel / @ref arm_compute::CLL2NormalizeLayer
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	351	- @ref arm_compute::CLQuantizationLayerKernel @ref arm_compute::CLMinMaxLayerKernel / @ref arm_compute::CLQuantizationLayer
				352	- @ref arm_compute::CLROIPoolingLayerKernel / @ref arm_compute::CLROIPoolingLayer
				353	- @ref arm_compute::CLReductionOperationKernel / @ref arm_compute::CLReductionOperation
				354	- @ref arm_compute::CLReshapeLayerKernel / @ref arm_compute::CLReshapeLayer
				355
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	356	v17.06 Public major release
				357	- Various bug fixes
				358	- Added support for fixed point 8 bit (QS8) to the various NEON machine learning kernels.
				359	- Added unit tests and benchmarks (AlexNet, LeNet)
				360	- Added support for sub tensors.
				361	- Added infrastructure to provide GPU specific optimisation for some OpenCL kernels.
				362	- Added @ref arm_compute::OMPScheduler (OpenMP) scheduler for NEON
				363	- Added @ref arm_compute::SingleThreadScheduler scheduler for NEON (For bare metal)
				364	- User can specify his own scheduler by implementing the @ref arm_compute::IScheduler interface.
				365	- New OpenCL kernels / functions:
				366	- @ref arm_compute::CLBatchNormalizationLayerKernel / @ref arm_compute::CLBatchNormalizationLayer
Giorgio Arena	04a8f8c	2017-11-23 11:45:24 +0000	[diff] [blame]	367	- @ref arm_compute::CLDepthConcatenateLayerKernel / @ref arm_compute::CLDepthConcatenateLayer
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	368	- @ref arm_compute::CLHOGOrientationBinningKernel @ref arm_compute::CLHOGBlockNormalizationKernel, @ref arm_compute::CLHOGDetectorKernel / @ref arm_compute::CLHOGDescriptor @ref arm_compute::CLHOGDetector @ref arm_compute::CLHOGGradient @ref arm_compute::CLHOGMultiDetection
				369	- @ref arm_compute::CLLocallyConnectedMatrixMultiplyKernel / @ref arm_compute::CLLocallyConnectedLayer
				370	- @ref arm_compute::CLWeightsReshapeKernel / @ref arm_compute::CLConvolutionLayerReshapeWeights
				371	- New C++ kernels:
				372	- @ref arm_compute::CPPDetectionWindowNonMaximaSuppressionKernel
				373	- New NEON kernels / functions:
				374	- @ref arm_compute::NEBatchNormalizationLayerKernel / @ref arm_compute::NEBatchNormalizationLayer
Giorgio Arena	04a8f8c	2017-11-23 11:45:24 +0000	[diff] [blame]	375	- @ref arm_compute::NEDepthConcatenateLayerKernel / @ref arm_compute::NEDepthConcatenateLayer
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	376	- @ref arm_compute::NEDirectConvolutionLayerKernel / @ref arm_compute::NEDirectConvolutionLayer
				377	- @ref arm_compute::NELocallyConnectedMatrixMultiplyKernel / @ref arm_compute::NELocallyConnectedLayer
				378	- @ref arm_compute::NEWeightsReshapeKernel / @ref arm_compute::NEConvolutionLayerReshapeWeights
				379
				380	v17.05 Public bug fixes release
				381	- Various bug fixes
				382	- Remaining of the functions ported to use accurate padding.
				383	- Library does not link against OpenCL anymore (It uses dlopen / dlsym at runtime instead to determine whether or not OpenCL is available).
				384	- Added "free" method to allocator.
				385	- Minimum version of g++ required for armv7 Linux changed from 4.8 to 4.9
				386
				387	v17.04 Public bug fixes release
				388
				389	The following functions have been ported to use the new accurate padding:
				390	- @ref arm_compute::CLColorConvertKernel
				391	- @ref arm_compute::CLEdgeNonMaxSuppressionKernel
				392	- @ref arm_compute::CLEdgeTraceKernel
				393	- @ref arm_compute::CLGaussianPyramidHorKernel
				394	- @ref arm_compute::CLGaussianPyramidVertKernel
				395	- @ref arm_compute::CLGradientKernel
				396	- @ref arm_compute::NEChannelCombineKernel
				397	- @ref arm_compute::NEFillArrayKernel
				398	- @ref arm_compute::NEGaussianPyramidHorKernel
				399	- @ref arm_compute::NEGaussianPyramidVertKernel
				400	- @ref arm_compute::NEHarrisScoreFP16Kernel
				401	- @ref arm_compute::NEHarrisScoreKernel
				402	- @ref arm_compute::NEHOGDetectorKernel
				403	- @ref arm_compute::NELogits1DMaxKernel
Diego Lopez Recas	35ceeb2	2017-12-04 18:56:10 +0000	[diff] [blame]	404	- arm_compute::NELogits1DShiftExpSumKernel
				405	- arm_compute::NELogits1DNormKernel
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	406	- @ref arm_compute::NENonMaximaSuppression3x3FP16Kernel
				407	- @ref arm_compute::NENonMaximaSuppression3x3Kernel
				408
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	409	v17.03.1 First Major public release of the sources
				410	- Renamed the library to arm_compute
				411	- New CPP target introduced for C++ kernels shared between NEON and CL functions.
				412	- New padding calculation interface introduced and ported most kernels / functions to use it.
				413	- New OpenCL kernels / functions:
Gian Marco	05288a2	2017-11-21 10:57:50 +0000	[diff] [blame]	414	- @ref arm_compute::CLGEMMLowpMatrixMultiplyKernel / arm_compute::CLGEMMLowp
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	415	- New NEON kernels / functions:
				416	- @ref arm_compute::NENormalizationLayerKernel / @ref arm_compute::NENormalizationLayer
				417	- @ref arm_compute::NETransposeKernel / @ref arm_compute::NETranspose
Diego Lopez Recas	35ceeb2	2017-12-04 18:56:10 +0000	[diff] [blame]	418	- @ref arm_compute::NELogits1DMaxKernel, arm_compute::NELogits1DShiftExpSumKernel, arm_compute::NELogits1DNormKernel / @ref arm_compute::NESoftmaxLayer
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	419	- @ref arm_compute::NEIm2ColKernel, @ref arm_compute::NECol2ImKernel, arm_compute::NEConvolutionLayerWeightsReshapeKernel / @ref arm_compute::NEConvolutionLayer
				420	- @ref arm_compute::NEGEMMMatrixAccumulateBiasesKernel / @ref arm_compute::NEFullyConnectedLayer
Gian Marco	e75a02b	2017-11-08 12:24:09 +0000	[diff] [blame]	421	- @ref arm_compute::NEGEMMLowpMatrixMultiplyKernel / arm_compute::NEGEMMLowp
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	422
				423	v17.03 Sources preview
				424	- New OpenCL kernels / functions:
				425	- @ref arm_compute::CLGradientKernel, @ref arm_compute::CLEdgeNonMaxSuppressionKernel, @ref arm_compute::CLEdgeTraceKernel / @ref arm_compute::CLCannyEdge
				426	- GEMM refactoring + FP16 support: @ref arm_compute::CLGEMMInterleave4x4Kernel, @ref arm_compute::CLGEMMTranspose1xWKernel, @ref arm_compute::CLGEMMMatrixMultiplyKernel, @ref arm_compute::CLGEMMMatrixAdditionKernel / @ref arm_compute::CLGEMM
				427	- @ref arm_compute::CLGEMMMatrixAccumulateBiasesKernel / @ref arm_compute::CLFullyConnectedLayer
				428	- @ref arm_compute::CLTransposeKernel / @ref arm_compute::CLTranspose
				429	- @ref arm_compute::CLLKTrackerInitKernel, @ref arm_compute::CLLKTrackerStage0Kernel, @ref arm_compute::CLLKTrackerStage1Kernel, @ref arm_compute::CLLKTrackerFinalizeKernel / @ref arm_compute::CLOpticalFlow
				430	- @ref arm_compute::CLNormalizationLayerKernel / @ref arm_compute::CLNormalizationLayer
				431	- @ref arm_compute::CLLaplacianPyramid, @ref arm_compute::CLLaplacianReconstruct
				432	- New NEON kernels / functions:
				433	- @ref arm_compute::NEActivationLayerKernel / @ref arm_compute::NEActivationLayer
				434	- GEMM refactoring + FP16 support (Requires armv8.2 CPU): @ref arm_compute::NEGEMMInterleave4x4Kernel, @ref arm_compute::NEGEMMTranspose1xWKernel, @ref arm_compute::NEGEMMMatrixMultiplyKernel, @ref arm_compute::NEGEMMMatrixAdditionKernel / @ref arm_compute::NEGEMM
				435	- @ref arm_compute::NEPoolingLayerKernel / @ref arm_compute::NEPoolingLayer
				436
				437	v17.02.1 Sources preview
				438	- New OpenCL kernels / functions:
				439	- @ref arm_compute::CLLogits1DMaxKernel, @ref arm_compute::CLLogits1DShiftExpSumKernel, @ref arm_compute::CLLogits1DNormKernel / @ref arm_compute::CLSoftmaxLayer
				440	- @ref arm_compute::CLPoolingLayerKernel / @ref arm_compute::CLPoolingLayer
Gian Marco Iodice	5cb4c42	2017-06-23 10:38:25 +0100	[diff] [blame]	441	- @ref arm_compute::CLIm2ColKernel, @ref arm_compute::CLCol2ImKernel, arm_compute::CLConvolutionLayerWeightsReshapeKernel / @ref arm_compute::CLConvolutionLayer
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	442	- @ref arm_compute::CLRemapKernel / @ref arm_compute::CLRemap
				443	- @ref arm_compute::CLGaussianPyramidHorKernel, @ref arm_compute::CLGaussianPyramidVertKernel / @ref arm_compute::CLGaussianPyramid, @ref arm_compute::CLGaussianPyramidHalf, @ref arm_compute::CLGaussianPyramidOrb
				444	- @ref arm_compute::CLMinMaxKernel, @ref arm_compute::CLMinMaxLocationKernel / @ref arm_compute::CLMinMaxLocation
				445	- @ref arm_compute::CLNonLinearFilterKernel / @ref arm_compute::CLNonLinearFilter
				446	- New NEON FP16 kernels (Requires armv8.2 CPU)
				447	- @ref arm_compute::NEAccumulateWeightedFP16Kernel
				448	- @ref arm_compute::NEBox3x3FP16Kernel
				449	- @ref arm_compute::NENonMaximaSuppression3x3FP16Kernel
				450
				451	v17.02 Sources preview
				452	- New OpenCL kernels / functions:
				453	- @ref arm_compute::CLActivationLayerKernel / @ref arm_compute::CLActivationLayer
				454	- @ref arm_compute::CLChannelCombineKernel / @ref arm_compute::CLChannelCombine
				455	- @ref arm_compute::CLDerivativeKernel / @ref arm_compute::CLChannelExtract
				456	- @ref arm_compute::CLFastCornersKernel / @ref arm_compute::CLFastCorners
				457	- @ref arm_compute::CLMeanStdDevKernel / @ref arm_compute::CLMeanStdDev
				458	- New NEON kernels / functions:
				459	- HOG / SVM: @ref arm_compute::NEHOGOrientationBinningKernel, @ref arm_compute::NEHOGBlockNormalizationKernel, @ref arm_compute::NEHOGDetectorKernel, arm_compute::NEHOGNonMaximaSuppressionKernel / @ref arm_compute::NEHOGDescriptor, @ref arm_compute::NEHOGDetector, @ref arm_compute::NEHOGGradient, @ref arm_compute::NEHOGMultiDetection
				460	- @ref arm_compute::NENonLinearFilterKernel / @ref arm_compute::NENonLinearFilter
				461	- Introduced a CLScheduler to manage the default context and command queue used by the runtime library and create synchronisation events.
				462	- Switched all the kernels / functions to use tensors instead of images.
				463	- Updated documentation to include instructions to build the library from sources.
				464
				465	v16.12 Binary preview release
				466	- Original release
				467
				468	@section S3_how_to_build How to build the library and the examples
				469
				470	@subsection S3_1_build_options Build options
				471
				472	scons 2.3 or above is required to build the library.
				473	To see the build options available simply run ```scons -h```:
				474
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	475	debug: Debug (yes\|no)
				476	default: False
				477	actual: False
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	478
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	479	asserts: Enable asserts (this flag is forced to 1 for debug=1) (yes\|no)
				480	default: False
				481	actual: False
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	482
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	483	arch: Target Architecture (armv7a\|arm64-v8a\|arm64-v8.2-a\|x86_32\|x86_64)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	484	default: armv7a
				485	actual: armv7a
				486
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	487	os: Target OS (linux\|android\|bare_metal)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	488	default: linux
				489	actual: linux
				490
Anthony Barbier	2d0ce77	2018-02-21 15:35:36 +0000	[diff] [blame]	491	build: Build type (native\|cross_compile\|embed_only)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	492	default: cross_compile
				493	actual: cross_compile
				494
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	495	examples: Build example programs (yes\|no)
				496	default: True
				497	actual: True
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	498
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	499	Werror: Enable/disable the -Werror compilation flag (yes\|no)
				500	default: True
				501	actual: True
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	502
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	503	opencl: Enable OpenCL support (yes\|no)
				504	default: True
				505	actual: True
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	506
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	507	neon: Enable Neon support (yes\|no)
				508	default: False
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	509	actual: False
				510
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	511	gles_compute: Enable OpenGL ES Compute Shader support (yes\|no)
				512	default: False
				513	actual: False
				514
				515	embed_kernels: Embed OpenCL kernels and OpenGL ES compute shader in library binary (yes\|no)
Anthony Barbier	cc0a80b	2017-12-15 11:37:29 +0000	[diff] [blame]	516	default: True
				517	actual: True
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	518
				519	set_soname: Set the library's soname and shlibversion (requires SCons 2.4 or above) (yes\|no)
				520	default: False
				521	actual: False
				522
				523	openmp: Enable OpenMP backend (yes\|no)
				524	default: False
				525	actual: False
				526
				527	cppthreads: Enable C++11 threads backend (yes\|no)
				528	default: True
				529	actual: True
				530
				531	build_dir: Specify sub-folder for the build ( /path/to/build_dir )
				532	default: .
				533	actual: .
				534
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	535	extra_cxx_flags: Extra CXX flags to be appended to the build command
				536	default:
				537	actual:
				538
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	539	pmu: Enable PMU counters (yes\|no)
				540	default: False
				541	actual: False
				542
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	543	mali: Enable Mali hardware counters (yes\|no)
				544	default: False
				545	actual: False
				546
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	547	validation_tests: Build validation test programs (yes\|no)
				548	default: False
				549	actual: False
				550
				551	benchmark_tests: Build benchmark test programs (yes\|no)
				552	default: False
				553	actual: False
				554
				555	@b debug / @b asserts:
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	556	- With debug=1 asserts are enabled, and the library is built with symbols and no optimisations enabled.
				557	- With debug=0 and asserts=1: Optimisations are enabled and symbols are removed, however all the asserts are still present (This is about 20% slower than the release build)
				558	- With debug=0 and asserts=0: All optimisations are enable and no validation is performed, if the application misuses the library it is likely to result in a crash. (Only use this mode once you are sure your application is working as expected).
				559
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	560	@b arch: The x86_32 and x86_64 targets can only be used with neon=0 and opencl=1.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	561
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	562	@b os: Choose the operating system you are targeting: Linux, Android or bare metal.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	563	@note bare metal can only be used for NEON (not OpenCL), only static libraries get built and NEON's multi-threading support is disabled.
				564
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	565	@b build: you can either build directly on your device (native) or cross compile from your desktop machine (cross-compile). In both cases make sure the compiler is available in your path.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	566
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	567	@note If you want to natively compile for 32bit on a 64bit ARM device running a 64bit OS then you will have to use cross-compile too.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	568
Anthony Barbier	2d0ce77	2018-02-21 15:35:36 +0000	[diff] [blame]	569	There is also an 'embed_only' option which will generate all the .embed files for the OpenCL kernels and / or OpenGLES compute shaders. This might be useful if using a different build system to compile the library.
				570
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	571	@b Werror: If you are compiling using the same toolchains as the ones used in this guide then there shouldn't be any warning and therefore you should be able to keep Werror=1. If with a different compiler version the library fails to build because of warnings interpreted as errors then, if you are sure the warnings are not important, you might want to try to build with Werror=0 (But please do report the issue either on Github or by an email to developer@arm.com so that the issue can be addressed).
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	572
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	573	@b opencl / @b neon / @b gles_compute: Choose which SIMD technology you want to target. (NEON for ARM Cortex-A CPUs or OpenCL / GLES_COMPUTE for ARM Mali GPUs)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	574
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	575	@b embed_kernels: For OpenCL / GLES_COMPUTE only: set embed_kernels=1 if you want the OpenCL / GLES_COMPUTE kernels to be built in the library's binaries instead of being read from separate ".cl" / ".cs" files. If embed_kernels is set to 0 then the application can set the path to the folder containing the OpenCL / GLES_COMPUTE kernel files by calling CLKernelLibrary::init() / GCKernelLibrary::init(). By default the path is set to "./cl_kernels" / "./cs_shaders".
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	576
				577	@b set_soname: Do you want to build the versioned version of the library ?
				578
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	579	If enabled the library will contain a SONAME and SHLIBVERSION and some symlinks will automatically be created between the objects.
				580	Example:
				581	libarm_compute_core.so -> libarm_compute_core.so.1.0.0
				582	libarm_compute_core.so.1 -> libarm_compute_core.so.1.0.0
				583	libarm_compute_core.so.1.0.0
				584
				585	@note This options is disabled by default as it requires SCons version 2.4 or above.
				586
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	587	@b extra_cxx_flags: Custom CXX flags which will be appended to the end of the build command.
				588
				589	@b build_dir: Build the library in a subfolder of the "build" folder. (Allows to build several configurations in parallel).
				590
				591	@b examples: Build or not the examples
				592
				593	@b validation_tests: Enable the build of the validation suite.
				594
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	595	@b benchmark_tests: Enable the build of the benchmark tests
				596
				597	@b pmu: Enable the PMU cycle counter to measure execution time in benchmark tests. (Your device needs to support it)
				598
Anthony Barbier	6a5627a	2017-09-26 14:42:02 +0100	[diff] [blame]	599	@b mali: Enable the collection of Mali hardware counters to measure execution time in benchmark tests. (Your device needs to have a Mali driver that supports it)
				600
Anthony Barbier	79c6178	2017-06-23 11:48:24 +0100	[diff] [blame]	601	@b openmp Build in the OpenMP scheduler for NEON.
				602
				603	@note Only works when building with g++ not clang++
				604
				605	@b cppthreads Build in the C++11 scheduler for NEON.
				606
				607	@sa arm_compute::Scheduler::set
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	608
Moritz Pflanzer	07674de	2017-07-21 09:39:36 +0100	[diff] [blame]	609	@subsection S3_2_linux Building for Linux
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	610
				611	@subsubsection S3_2_1_library How to build the library ?
				612
				613	For Linux, the library was successfully built and tested using the following Linaro GCC toolchain:
				614
				615	- gcc-linaro-arm-linux-gnueabihf-4.9-2014.07_linux
				616	- gcc-linaro-4.9-2016.02-x86_64_aarch64-linux-gnu
				617	- gcc-linaro-6.3.1-2017.02-i686_aarch64-linux-gnu
				618
				619	@note If you are building with opencl=1 then scons will expect to find libOpenCL.so either in the current directory or in "build" (See the section below if you need a stub OpenCL library to link against)
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	620	@note If you are building with gles_compute=1 then scons will expect to find libEGL.so / libGLESv1_CM.so / libGLESv2.so either in the current directory or in "build" (See the section below if you need a stub OpenCL library to link against)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	621
				622	To cross-compile the library in debug mode, with NEON only support, for Linux 32bit:
				623
				624	scons Werror=1 -j8 debug=1 neon=1 opencl=0 os=linux arch=armv7a
				625
				626	To cross-compile the library in asserts mode, with OpenCL only support, for Linux 64bit:
				627
				628	scons Werror=1 -j8 debug=0 asserts=1 neon=0 opencl=1 embed_kernels=1 os=linux arch=arm64-v8a
				629
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	630	To cross-compile the library in asserts mode, with GLES_COMPUTE only support, for Linux 64bit:
				631
				632	scons Werror=1 -j8 debug=0 asserts=1 neon=0 opencl=0 gles_compute=1 embed_kernels=1 os=linux arch=arm64-v8a
				633
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	634	You can also compile the library natively on an ARM device by using <b>build=native</b>:
				635
				636	scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=arm64-v8a build=native
				637	scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=armv7a build=native
				638
				639	@note g++ for ARM is mono-arch, therefore if you want to compile for Linux 32bit on a Linux 64bit platform you will have to use a cross compiler.
				640
				641	For example on a 64bit Debian based system you would have to install <b>g++-arm-linux-gnueabihf</b>
				642
				643	apt-get install g++-arm-linux-gnueabihf
				644
				645	Then run
				646
				647	scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=armv7a build=cross_compile
				648
				649	or simply remove the build parameter as build=cross_compile is the default value:
				650
				651	scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=armv7a
				652
				653	@attention To cross compile with opencl=1 you need to make sure to have a version of libOpenCL matching your target architecture.
				654
				655	@subsubsection S3_2_2_examples How to manually build the examples ?
				656
				657	The examples get automatically built by scons as part of the build process of the library described above. This section just describes how you can build and link your own application against our library.
				658
				659	@note The following command lines assume the arm_compute and libOpenCL binaries are present in the current directory or in the system library path. If this is not the case you can specify the location of the pre-built library with the compiler option -L. When building the OpenCL example the commands below assume that the CL headers are located in the include folder where the command is executed.
				660
				661	To cross compile a NEON example for Linux 32bit:
				662
Anthony Barbier	b2881fc	2017-09-29 17:12:12 +0100	[diff] [blame]	663	arm-linux-gnueabihf-g++ examples/neon_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute -larm_compute_core -o neon_convolution
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	664
				665	To cross compile a NEON example for Linux 64bit:
				666
Anthony Barbier	b2881fc	2017-09-29 17:12:12 +0100	[diff] [blame]	667	aarch64-linux-gnu-g++ examples/neon_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -L. -larm_compute -larm_compute_core -o neon_convolution
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	668
				669	(notice the only difference with the 32 bit command is that we don't need the -mfpu option and the compiler's name is different)
				670
				671	To cross compile an OpenCL example for Linux 32bit:
				672
Anthony Barbier	b2881fc	2017-09-29 17:12:12 +0100	[diff] [blame]	673	arm-linux-gnueabihf-g++ examples/cl_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute -larm_compute_core -lOpenCL -o cl_convolution -DARM_COMPUTE_CL
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	674
				675	To cross compile an OpenCL example for Linux 64bit:
				676
Anthony Barbier	b2881fc	2017-09-29 17:12:12 +0100	[diff] [blame]	677	aarch64-linux-gnu-g++ examples/cl_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -L. -larm_compute -larm_compute_core -lOpenCL -o cl_convolution -DARM_COMPUTE_CL
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	678
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	679	To cross compile a GLES example for Linux 32bit:
				680
				681	arm-linux-gnueabihf-g++ examples/gc_absdiff.cpp utils/Utils.cpp -I. -Iinclude/ -L. -larm_compute -larm_compute_core -std=c++11 -mfpu=neon -DARM_COMPUTE_GC -Iinclude/linux/ -o gc_absdiff
				682
				683	To cross compile a GLES example for Linux 64bit:
				684
				685	aarch64-linux-gnu-g++ examples/gc_absdiff.cpp utils/Utils.cpp -I. -Iinclude/ -L. -larm_compute -larm_compute_core -std=c++11 -DARM_COMPUTE_GC -Iinclude/linux/ -o gc_absdiff
				686
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	687	(notice the only difference with the 32 bit command is that we don't need the -mfpu option and the compiler's name is different)
				688
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	689	To cross compile the examples with the Graph API, such as graph_lenet.cpp, you need to link the examples against arm_compute_graph.so too.
				690
				691	@note The compute library must currently be built with both neon and opencl enabled - neon=1 and opencl=1
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	692
				693	i.e. to cross compile the "graph_lenet" example for Linux 32bit:
				694
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	695	arm-linux-gnueabihf-g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	696
				697	i.e. to cross compile the "graph_lenet" example for Linux 64bit:
				698
Isabella Gottardi	b28f29d	2017-11-09 17:05:07 +0000	[diff] [blame]	699	aarch64-linux-gnu-g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	700
				701	(notice the only difference with the 32 bit command is that we don't need the -mfpu option and the compiler's name is different)
				702
Anthony Barbier	e500747	2017-10-27 15:01:44 +0100	[diff] [blame]	703	@note If compiling using static libraries, this order must be followed when linking: arm_compute_graph_static, arm_compute, arm_compute_core
				704
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	705	To compile natively (i.e directly on an ARM device) for NEON for Linux 32bit:
				706
Anthony Barbier	b2881fc	2017-09-29 17:12:12 +0100	[diff] [blame]	707	g++ examples/neon_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -larm_compute -larm_compute_core -o neon_convolution
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	708
				709	To compile natively (i.e directly on an ARM device) for NEON for Linux 64bit:
				710
Anthony Barbier	b2881fc	2017-09-29 17:12:12 +0100	[diff] [blame]	711	g++ examples/neon_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute -larm_compute_core -o neon_convolution
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	712
				713	(notice the only difference with the 32 bit command is that we don't need the -mfpu option)
				714
				715	To compile natively (i.e directly on an ARM device) for OpenCL for Linux 32bit or Linux 64bit:
				716
Anthony Barbier	b2881fc	2017-09-29 17:12:12 +0100	[diff] [blame]	717	g++ examples/cl_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute -larm_compute_core -lOpenCL -o cl_convolution -DARM_COMPUTE_CL
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	718
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	719	To compile natively (i.e directly on an ARM device) for GLES for Linux 32bit or Linux 64bit:
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	720
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	721	g++ examples/gc_absdiff.cpp utils/Utils.cpp -I. -Iinclude/ -L. -larm_compute -larm_compute_core -std=c++11 -DARM_COMPUTE_GC -Iinclude/linux/ -o gc_absdiff
				722
				723	To compile natively the examples with the Graph API, such as graph_lenet.cpp, you need to link the examples against arm_compute_graph.so too.
				724	@note The compute library must currently be built with both neon and opencl enabled - neon=1 and opencl=1
				725
				726	i.e. to natively compile the "graph_lenet" example for Linux 32bit:
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	727
Isabella Gottardi	b28f29d	2017-11-09 17:05:07 +0000	[diff] [blame]	728	g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	729
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	730	i.e. to natively compile the "graph_lenet" example for Linux 64bit:
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	731
Isabella Gottardi	b28f29d	2017-11-09 17:05:07 +0000	[diff] [blame]	732	g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	733
				734	(notice the only difference with the 32 bit command is that we don't need the -mfpu option)
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	735
Anthony Barbier	e500747	2017-10-27 15:01:44 +0100	[diff] [blame]	736	@note If compiling using static libraries, this order must be followed when linking: arm_compute_graph_static, arm_compute, arm_compute_core
				737
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	738	@note These two commands assume libarm_compute.so is available in your library path, if not add the path to it using -L
				739
				740	To run the built executable simply run:
				741
				742	LD_LIBRARY_PATH=build ./neon_convolution
				743
				744	or
				745
				746	LD_LIBRARY_PATH=build ./cl_convolution
				747
Moritz Pflanzer	07674de	2017-07-21 09:39:36 +0100	[diff] [blame]	748	@subsection S3_3_android Building for Android
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	749
				750	For Android, the library was successfully built and tested using Google's standalone toolchains:
Anthony Barbier	a8a28f6	2018-02-26 19:16:32 +0000	[diff] [blame]	751	- clang++ from NDK r16b for armv7a
				752	- clang++ from NDK r16b for arm64-v8a
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	753
				754	Here is a guide to <a href="https://developer.android.com/ndk/guides/standalone_toolchain.html">create your Android standalone toolchains from the NDK</a>
				755
Anthony Barbier	a8a28f6	2018-02-26 19:16:32 +0000	[diff] [blame]	756	- Download the NDK r16b from here: https://developer.android.com/ndk/downloads/index.html
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	757	- Make sure you have Python 2 installed on your machine.
				758	- Generate the 32 and/or 64 toolchains by running the following commands:
				759
Anthony Barbier	a8a28f6	2018-02-26 19:16:32 +0000	[diff] [blame]	760	$NDK/build/tools/make_standalone_toolchain.py --arch arm64 --install-dir $MY_TOOLCHAINS/aarch64-linux-android-ndk-r16b --stl gnustl --api 21
				761	$NDK/build/tools/make_standalone_toolchain.py --arch arm --install-dir $MY_TOOLCHAINS/arm-linux-android-ndk-r16b --stl gnustl --api 21
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	762
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	763	@attention Due to some NDK issues make sure you use clang++ & gnustl
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	764
				765	@note Make sure to add the toolchains to your PATH: export PATH=$PATH:$MY_TOOLCHAINS/aarch64-linux-android-4.9/bin:$MY_TOOLCHAINS/arm-linux-androideabi-4.9/bin
				766
				767	@subsubsection S3_3_1_library How to build the library ?
				768
				769	@note If you are building with opencl=1 then scons will expect to find libOpenCL.so either in the current directory or in "build" (See the section below if you need a stub OpenCL library to link against)
				770
				771	To cross-compile the library in debug mode, with NEON only support, for Android 32bit:
				772
				773	CXX=clang++ CC=clang scons Werror=1 -j8 debug=1 neon=1 opencl=0 os=android arch=armv7a
				774
				775	To cross-compile the library in asserts mode, with OpenCL only support, for Android 64bit:
				776
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	777	CXX=clang++ CC=clang scons Werror=1 -j8 debug=0 asserts=1 neon=0 opencl=1 embed_kernels=1 os=android arch=arm64-v8a
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	778
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	779	To cross-compile the library in asserts mode, with GLES_COMPUTE only support, for Android 64bit:
				780
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	781	CXX=clang++ CC=clang scons Werror=1 -j8 debug=0 asserts=1 neon=0 opencl=0 gles_compute=1 embed_kernels=1 os=android arch=arm64-v8a
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	782
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	783	@subsubsection S3_3_2_examples How to manually build the examples ?
				784
				785	The examples get automatically built by scons as part of the build process of the library described above. This section just describes how you can build and link your own application against our library.
				786
Anthony Barbier	fabb038	2017-06-23 14:42:52 +0100	[diff] [blame]	787	@note The following command lines assume the arm_compute and libOpenCL binaries are present in the current directory or in the system library path. If this is not the case you can specify the location of the pre-built library with the compiler option -L. When building the OpenCL example the commands below assume that the CL headers are located in the include folder where the command is executed.
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	788
				789	Once you've got your Android standalone toolchain built and added to your path you can do the following:
				790
				791	To cross compile a NEON example:
				792
				793	#32 bit:
Georgios Pinitas	9873ea3	2017-12-05 15:28:55 +0000	[diff] [blame]	794	arm-linux-androideabi-clang++ examples/neon_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute-static -larm_compute_core-static -L. -o neon_convolution_arm -static-libstdc++ -pie
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	795	#64 bit:
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	796	aarch64-linux-android-clang++ examples/neon_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute-static -larm_compute_core-static -L. -o neon_convolution_aarch64 -static-libstdc++ -pie
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	797
				798	To cross compile an OpenCL example:
				799
				800	#32 bit:
Georgios Pinitas	9873ea3	2017-12-05 15:28:55 +0000	[diff] [blame]	801	arm-linux-androideabi-clang++ examples/cl_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute-static -larm_compute_core-static -L. -o cl_convolution_arm -static-libstdc++ -pie -lOpenCL -DARM_COMPUTE_CL
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	802	#64 bit:
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	803	aarch64-linux-android-clang++ examples/cl_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute-static -larm_compute_core-static -L. -o cl_convolution_aarch64 -static-libstdc++ -pie -lOpenCL -DARM_COMPUTE_CL
				804
				805	To cross compile a GLES example:
Anthony Barbier	cc0a80b	2017-12-15 11:37:29 +0000	[diff] [blame]	806
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	807	#32 bit:
				808	arm-linux-androideabi-clang++ examples/gc_absdiff.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute-static -larm_compute_core-static -L. -o gc_absdiff_arm -static-libstdc++ -pie -DARM_COMPUTE_GC
				809	#64 bit:
				810	aarch64-linux-android-clang++ examples/gc_absdiff.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -larm_compute-static -larm_compute_core-static -L. -o gc_absdiff_aarch64 -static-libstdc++ -pie -DARM_COMPUTE_GC
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	811
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	812	To cross compile the examples with the Graph API, such as graph_lenet.cpp, you need to link the library arm_compute_graph also.
				813	(notice the compute library has to be built with both neon and opencl enabled - neon=1 and opencl=1)
				814
				815	#32 bit:
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	816	arm-linux-androideabi-clang++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -Wl,--whole-archive -larm_compute_graph-static -Wl,--no-whole-archive -larm_compute-static -larm_compute_core-static -L. -o graph_lenet_arm -static-libstdc++ -pie -lOpenCL -DARM_COMPUTE_CL
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	817	#64 bit:
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	818	aarch64-linux-android-clang++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -Wl,--whole-archive -larm_compute_graph-static -Wl,--no-whole-archive -larm_compute-static -larm_compute_core-static -L. -o graph_lenet_aarch64 -static-libstdc++ -pie -lOpenCL -DARM_COMPUTE_CL
Gian Marco Iodice	daec1aa	2017-09-29 12:03:18 +0100	[diff] [blame]	819
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	820	@note Due to some issues in older versions of the Mali OpenCL DDK (<= r13p0), we recommend to link arm_compute statically on Android.
Anthony Barbier	20dbb82	2017-12-13 21:19:39 +0000	[diff] [blame]	821	@note When linked statically the arm_compute_graph library currently needs the --whole-archive linker flag in order to work properly
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	822
				823	Then you need to do is upload the executable and the shared library to the device using ADB:
				824
				825	adb push neon_convolution_arm /data/local/tmp/
				826	adb push cl_convolution_arm /data/local/tmp/
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	827	adb push gc_absdiff_arm /data/local/tmp/
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	828	adb shell chmod 777 -R /data/local/tmp/
				829
				830	And finally to run the example:
				831
				832	adb shell /data/local/tmp/neon_convolution_arm
				833	adb shell /data/local/tmp/cl_convolution_arm
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	834	adb shell /data/local/tmp/gc_absdiff_arm
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	835
				836	For 64bit:
				837
				838	adb push neon_convolution_aarch64 /data/local/tmp/
				839	adb push cl_convolution_aarch64 /data/local/tmp/
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	840	adb push gc_absdiff_aarch64 /data/local/tmp/
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	841	adb shell chmod 777 -R /data/local/tmp/
				842
				843	And finally to run the example:
				844
				845	adb shell /data/local/tmp/neon_convolution_aarch64
				846	adb shell /data/local/tmp/cl_convolution_aarch64
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	847	adb shell /data/local/tmp/gc_absdiff_aarch64
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	848
Michalis Spyrou	6e52ba3	2017-10-04 15:40:38 +0100	[diff] [blame]	849	@subsection S3_4_bare_metal Building for bare metal
				850
				851	For bare metal, the library was successfully built using linaros's latest (gcc-linaro-6.3.1-2017.05) bare metal toolchains:
				852	- arm-eabi for armv7a
				853	- aarch64-elf for arm64-v8a
				854
				855	Download linaro for <a href="https://releases.linaro.org/components/toolchain/binaries/6.3-2017.05/arm-eabi/">armv7a</a> and <a href="https://releases.linaro.org/components/toolchain/binaries/6.3-2017.05/aarch64-elf/">arm64-v8a</a>.
				856
				857	@note Make sure to add the toolchains to your PATH: export PATH=$PATH:$MY_TOOLCHAINS/gcc-linaro-6.3.1-2017.05-x86_64_aarch64-elf/bin:$MY_TOOLCHAINS/gcc-linaro-6.3.1-2017.05-x86_64_arm-eabi/bin
				858
				859	@subsubsection S3_4_1_library How to build the library ?
				860
				861	To cross-compile the library with NEON support for baremetal arm64-v8a:
				862
				863	scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=bare_metal arch=arm64-v8a build=cross_compile cppthreads=0 openmp=0 standalone=1
				864
				865	@subsubsection S3_4_2_examples How to manually build the examples ?
				866
				867	Examples are disabled when building for bare metal. If you want to build the examples you need to provide a custom bootcode depending on the target architecture and link against the compute library. More information about bare metal bootcode can be found <a href="http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dai0527a/index.html">here</a>.
				868
				869	@subsection S3_5_windows_host Building on a Windows host system
Moritz Pflanzer	07674de	2017-07-21 09:39:36 +0100	[diff] [blame]	870
				871	Using `scons` directly from the Windows command line is known to cause
				872	problems. The reason seems to be that if `scons` is setup for cross-compilation
				873	it gets confused about Windows style paths (using backslashes). Thus it is
				874	recommended to follow one of the options outlined below.
				875
Michalis Spyrou	6e52ba3	2017-10-04 15:40:38 +0100	[diff] [blame]	876	@subsubsection S3_5_1_ubuntu_on_windows Bash on Ubuntu on Windows
Moritz Pflanzer	07674de	2017-07-21 09:39:36 +0100	[diff] [blame]	877
				878	The best and easiest option is to use
				879	<a href="https://msdn.microsoft.com/en-gb/commandline/wsl/about">Ubuntu on Windows</a>.
				880	This feature is still marked as beta and thus might not be available.
				881	However, if it is building the library is as simple as opening a *Bash on
				882	Ubuntu on Windows* shell and following the general guidelines given above.
				883
Michalis Spyrou	6e52ba3	2017-10-04 15:40:38 +0100	[diff] [blame]	884	@subsubsection S3_5_2_cygwin Cygwin
Moritz Pflanzer	07674de	2017-07-21 09:39:36 +0100	[diff] [blame]	885
				886	If the Windows subsystem for Linux is not available <a href="https://www.cygwin.com/">Cygwin</a>
				887	can be used to install and run `scons`. In addition to the default packages
				888	installed by Cygwin `scons` has to be selected in the installer. (`git` might
				889	also be useful but is not strictly required if you already have got the source
				890	code of the library.) Linaro provides pre-built versions of
				891	<a href="http://releases.linaro.org/components/toolchain/binaries/">GCC cross-compilers</a>
				892	that can be used from the Cygwin terminal. When building for Android the
				893	compiler is included in the Android standalone toolchain. After everything has
				894	been set up in the Cygwin terminal the general guide on building the library
				895	can be followed.
				896
Michalis Spyrou	6e52ba3	2017-10-04 15:40:38 +0100	[diff] [blame]	897	@subsection S3_6_cl_stub_library The OpenCL stub library
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	898
				899	In the opencl-1.2-stubs folder you will find the sources to build a stub OpenCL library which then can be used to link your application or arm_compute against.
				900
				901	If you preferred you could retrieve the OpenCL library from your device and link against this one but often this library will have dependencies on a range of system libraries forcing you to link your application against those too even though it is not using them.
				902
				903	@warning This OpenCL library provided is a stub and not a real implementation. You can use it to resolve OpenCL's symbols in arm_compute while building the example but you must make sure the real libOpenCL.so is in your PATH when running the example or it will not work.
				904
				905	To cross-compile the stub OpenCL library simply run:
				906
				907	<target-prefix>-gcc -o libOpenCL.so -Iinclude opencl-1.2-stubs/opencl_stubs.c -fPIC -shared
				908
				909	For example:
				910
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	911	#Linux 32bit
				912	arm-linux-gnueabihf-gcc -o libOpenCL.so -Iinclude opencl-1.2-stubs/opencl_stubs.c -fPIC -shared
				913	#Linux 64bit
				914	aarch64-linux-gnu-gcc -o libOpenCL.so -Iinclude -shared opencl-1.2-stubs/opencl_stubs.c -fPIC
				915	#Android 32bit
				916	arm-linux-androideabi-clang -o libOpenCL.so -Iinclude -shared opencl-1.2-stubs/opencl_stubs.c -fPIC -shared
				917	#Android 64bit
Anthony Barbier	14c86a9	2017-12-14 16:27:41 +0000	[diff] [blame]	918	aarch64-linux-android-clang -o libOpenCL.so -Iinclude -shared opencl-1.2-stubs/opencl_stubs.c -fPIC -shared
				919
				920	@subsection S3_7_gles_stub_library The Linux OpenGLES and EGL stub libraries
				921
				922	In the opengles-3.1-stubs folder you will find the sources to build stub EGL and OpenGLES libraries which then can be used to link your Linux application of arm_compute against.
				923
				924	@note The stub libraries are only needed on Linux. For Android, the NDK toolchains already provide the meta-EGL and meta-GLES libraries.
				925
				926	To cross-compile the stub OpenGLES and EGL libraries simply run:
				927
				928	<target-prefix>-gcc -o libEGL.so -Iinclude/linux opengles-3.1-stubs/EGL.c -fPIC -shared
				929	<target-prefix>-gcc -o libGLESv2.so -Iinclude/linux opengles-3.1-stubs/GLESv2.c -fPIC -shared
				930
				931	#Linux 32bit
				932	arm-linux-gnueabihf-gcc -o libEGL.so -Iinclude/linux opengles-3.1-stubs/EGL.c -fPIC -shared
				933	arm-linux-gnueabihf-gcc -o libGLESv2.so -Iinclude/linux opengles-3.1-stubs/GLESv2.c -fPIC -shared
				934
				935	#Linux 64bit
				936	aarch64-linux-gnu-gcc -o libEGL.so -Iinclude/linux opengles-3.1-stubs/EGL.c -fPIC -shared
				937	aarch64-linux-gnu-gcc -o libGLESv2.so -Iinclude/linux opengles-3.1-stubs/GLESv2.c -fPIC -shared
Anthony Barbier	6ff3b19	2017-09-04 18:44:23 +0100	[diff] [blame]	938	*/