1/// 2/// Copyright (c) 2017-2023 Arm Limited. 3/// 4/// SPDX-License-Identifier: MIT 5/// 6/// Permission is hereby granted, free of charge, to any person obtaining a copy 7/// of this software and associated documentation files (the "Software"), to 8/// deal in the Software without restriction, including without limitation the 9/// rights to use, copy, modify, merge, publish, distribute, sublicense, and/or 10/// sell copies of the Software, and to permit persons to whom the Software is 11/// furnished to do so, subject to the following conditions: 12/// 13/// The above copyright notice and this permission notice shall be included in all 14/// copies or substantial portions of the Software. 15/// 16/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 17/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 18/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 19/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 20/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 21/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 22/// SOFTWARE. 23/// 24namespace arm_compute 25{ 26/** @page how_to_build How to Build and Run Examples 27 28@tableofcontents 29 30@section S1_1_build_options Build options 31 32scons 2.3 or above is required to build the library. 33To see the build options available simply run ```scons -h``` 34 35@section S1_2_linux Building for Linux 36 37@subsection S1_2_1_library How to build the library ? 38 39For Linux, the library was successfully built and tested using the following Linaro GCC toolchain: 40 41 - gcc-linaro-6.3.1-2017.05-x86_64_arm-linux-gnueabihf 42 - gcc-linaro-6.3.1-2017.05-x86_64_aarch64-linux-gnu 43 44To cross-compile the library in debug mode, with Arm® Neon™ only support, for Linux 32bit: 45 46 scons Werror=1 -j8 debug=1 neon=1 opencl=0 os=linux arch=armv7a 47 48To cross-compile the library in asserts mode, with OpenCL only support, for Linux 64bit: 49 50 scons Werror=1 -j8 debug=0 asserts=1 neon=0 opencl=1 embed_kernels=1 os=linux arch=armv8a 51 52You can also compile the library natively on an Arm device by using <b>build=native</b>: 53 54 scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=armv8a build=native 55 scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=armv7a build=native 56 57@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. 58 59For example on a 64bit Debian based system you would have to install <b>g++-arm-linux-gnueabihf</b> 60 61 apt-get install g++-arm-linux-gnueabihf 62 63Then run 64 65 scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=armv7a build=cross_compile 66 67or simply remove the build parameter as build=cross_compile is the default value: 68 69 scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=linux arch=armv7a 70 71@subsection S1_2_2_examples How to manually build the examples ? 72 73The 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. 74 75@note The following command lines assume the arm_compute libraries 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 libraries 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. 76 77To cross compile a Arm® Neon™ example for Linux 32bit: 78 79 arm-linux-gnueabihf-g++ examples/neon_cnn.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -mfpu=neon -L. -larm_compute -larm_compute_core -o neon_cnn 80 81To cross compile a Arm® Neon™ example for Linux 64bit: 82 83 aarch64-linux-gnu-g++ examples/neon_cnn.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -L. -larm_compute -larm_compute_core -o neon_cnn 84 85(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) 86 87To cross compile an OpenCL example for Linux 32bit: 88 89 arm-linux-gnueabihf-g++ examples/cl_sgemm.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -mfpu=neon -L. -larm_compute -larm_compute_core -o cl_sgemm -DARM_COMPUTE_CL 90 91To cross compile an OpenCL example for Linux 64bit: 92 93 aarch64-linux-gnu-g++ examples/cl_sgemm.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -L. -larm_compute -larm_compute_core -o cl_sgemm -DARM_COMPUTE_CL 94 95(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) 96 97To 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. 98 99i.e. to cross compile the "graph_lenet" example for Linux 32bit: 100 101 arm-linux-gnueabihf-g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp utils/CommonGraphOptions.cpp -I. -Iinclude -std=c++14 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet 102 103i.e. to cross compile the "graph_lenet" example for Linux 64bit: 104 105 aarch64-linux-gnu-g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp utils/CommonGraphOptions.cpp -I. -Iinclude -std=c++14 -L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet 106 107(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) 108 109@note If compiling using static libraries, this order must be followed when linking: arm_compute_graph_static, arm_compute, arm_compute_core 110 111To compile natively (i.e directly on an Arm device) for Arm® Neon™ for Linux 32bit: 112 113 g++ examples/neon_cnn.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -mfpu=neon -larm_compute -larm_compute_core -o neon_cnn 114 115To compile natively (i.e directly on an Arm device) for Arm® Neon™ for Linux 64bit: 116 117 g++ examples/neon_cnn.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -larm_compute -larm_compute_core -o neon_cnn 118 119(notice the only difference with the 32 bit command is that we don't need the -mfpu option) 120 121To compile natively (i.e directly on an Arm device) for OpenCL for Linux 32bit or Linux 64bit: 122 123 g++ examples/cl_sgemm.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -larm_compute -larm_compute_core -o cl_sgemm -DARM_COMPUTE_CL 124 125To 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. 126 127i.e. to natively compile the "graph_lenet" example for Linux 32bit: 128 129 g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp utils/CommonGraphOptions.cpp -I. -Iinclude -std=c++14 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet 130 131i.e. to natively compile the "graph_lenet" example for Linux 64bit: 132 133 g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp utils/CommonGraphOptions.cpp -I. -Iinclude -std=c++14 -L. -larm_compute_graph -larm_compute -larm_compute_core -Wl,--allow-shlib-undefined -o graph_lenet 134 135(notice the only difference with the 32 bit command is that we don't need the -mfpu option) 136 137@note If compiling using static libraries, this order must be followed when linking: arm_compute_graph_static, arm_compute, arm_compute_core 138 139@note These two commands assume libarm_compute.so is available in your library path, if not add the path to it using -L (e.g. -Llib/linux-armv8a-neon-cl-asserts/) 140@note You might need to export the path to OpenCL library as well in your LD_LIBRARY_PATH if Compute Library was built with OpenCL enabled. 141 142To run the built executable simply run: 143 144 LD_LIBRARY_PATH=build ./neon_cnn 145 146or 147 148 LD_LIBRARY_PATH=build ./cl_sgemm 149 150@note Examples accept different types of arguments, to find out what they are run the example with \a --help as an argument. If no arguments are specified then random values will be used to execute the graph. 151 152For example: 153 154 LD_LIBRARY_PATH=. ./graph_lenet --help 155 156Below is a list of the common parameters among the graph examples : 157@snippet utils/CommonGraphOptions.h Common graph examples parameters 158 159@subsection S1_2_3_sve Build for SVE or SVE2 160 161In order to build for SVE or SVE2 you need a compiler that supports them. You can find more information in the following these links: 162 -# GCC: https://developer.arm.com/tools-and-software/open-source-software/developer-tools/gnu-toolchain/sve-support 163 -# LLVM: https://developer.arm.com/tools-and-software/open-source-software/developer-tools/llvm-toolchain/sve-support 164 165@note You the need to indicate the toolchains using the scons "toolchain_prefix" parameter. 166 167An example build command with SVE is: 168 169 scons arch=armv8.2-a-sve os=linux build_dir=arm64 -j55 standalone=0 opencl=0 openmp=0 validation_tests=1 neon=1 cppthreads=1 toolchain_prefix=aarch64-none-linux-gnu- 170 171@subsection S1_2_4_sme Build for SME2 172 173In order to build for SME2 you need to use a compiler that supports SVE2 and enable SVE2 in the build as well. 174 175@note You the need to indicate the toolchains using the scons "toolchain_prefix" parameter. 176 177An example build command with SME2 is: 178 179 scons arch=armv8.6-a-sve2-sme2 os=linux build_dir=arm64 -j55 standalone=0 opencl=0 openmp=0 validation_tests=1 neon=1 cppthreads=1 toolchain_prefix=aarch64-none-linux-gnu- 180 181@section S1_3_android Building for Android 182 183For Android, the library was successfully built and tested using Google's standalone toolchains: 184 - clang++ from NDK r20b for armv8a 185 - clang++ from NDK r20b for armv8.2-a with FP16 support 186 187(From 23.02, NDK >= r20b is highly recommended) For NDK r18 or older, 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>: 188- Download the NDK r18b from here: https://developer.android.com/ndk/downloads/index.html to directory $NDK 189- Make sure you have Python 2.7 installed on your machine. 190- Generate the 32 and/or 64 toolchains by running the following commands to your toolchain directory $MY_TOOLCHAINS: 191 192 $NDK/build/tools/make_standalone_toolchain.py --arch arm64 --install-dir $MY_TOOLCHAINS/aarch64-linux-android-ndk-r18b --stl libc++ --api 21 193 194 $NDK/build/tools/make_standalone_toolchain.py --arch arm --install-dir $MY_TOOLCHAINS/arm-linux-android-ndk-r18b --stl libc++ --api 21 195 196For NDK r19 or newer, you can directly <a href="https://developer.android.com/ndk/downloads">Download</a> the NDK package for your development platform, without the need to launch the make_standalone_toolchain.py script. You can find all the prebuilt binaries inside $NDK/toolchains/llvm/prebuilt/$OS_ARCH/bin/. 197 198@parblock 199@attention The building script will look for a binary named "aarch64-linux-android-clang++", while the prebuilt binaries will have their API version as a suffix to their filename (e.g. "aarch64-linux-android21-clang++"). You can instruct scons to use the correct version by using a combination of the toolchain_prefix and the "CC" "CXX" environment variables. 200@attention For this particular example, you can specify: 201 202 CC=clang CXX=clang++ scons toolchain_prefix=aarch64-linux-android21- 203 204@attention or: 205 206 CC=aarch64-linux-android21-clang CXX=aarch64-linux-android21-clang++ scons toolchain_prefix="" 207 208@endparblock 209 210@parblock 211@attention We used to use gnustl but as of NDK r17 it is deprecated so we switched to libc++ 212@endparblock 213 214@note Make sure to add the toolchains to your PATH: 215 216 export PATH=$PATH:$MY_TOOLCHAINS/aarch64-linux-android-ndk-r18b/bin:$MY_TOOLCHAINS/arm-linux-android-ndk-r18b/bin 217 218@subsection S1_3_1_library How to build the library ? 219 220To cross-compile the library in debug mode, with Arm® Neon™ only support, for Android 32bit: 221 222 CXX=clang++ CC=clang scons Werror=1 -j8 debug=1 neon=1 opencl=0 os=android arch=armv7a 223 224To cross-compile the library in asserts mode, with OpenCL only support, for Android 64bit: 225 226 CXX=clang++ CC=clang scons Werror=1 -j8 debug=0 asserts=1 neon=0 opencl=1 embed_kernels=1 os=android arch=armv8a 227 228@subsection S1_3_2_examples How to manually build the examples ? 229 230The 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. 231 232@note The following command lines assume the arm_compute libraries 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 libraries 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. 233 234Once you've got your Android standalone toolchain built and added to your path you can do the following: 235 236To cross compile a Arm® Neon™ example: 237 238 #32 bit: 239 arm-linux-androideabi-clang++ examples/neon_cnn.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -larm_compute-static -larm_compute_core-static -L. -o neon_cnn_arm -static-libstdc++ -pie 240 #64 bit: 241 aarch64-linux-android-clang++ examples/neon_cnn.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -larm_compute-static -larm_compute_core-static -L. -o neon_cnn_aarch64 -static-libstdc++ -pie 242 243To cross compile an OpenCL example: 244 245 #32 bit: 246 arm-linux-androideabi-clang++ examples/cl_sgemm.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -larm_compute-static -larm_compute_core-static -L. -o cl_sgemm_arm -static-libstdc++ -pie -DARM_COMPUTE_CL 247 #64 bit: 248 aarch64-linux-android-clang++ examples/cl_sgemm.cpp utils/Utils.cpp -I. -Iinclude -std=c++14 -larm_compute-static -larm_compute_core-static -L. -o cl_sgemm_aarch64 -static-libstdc++ -pie -DARM_COMPUTE_CL 249 250To cross compile the examples with the Graph API, such as graph_lenet.cpp, you need to link the library arm_compute_graph also. 251 252 #32 bit: 253 arm-linux-androideabi-clang++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp utils/CommonGraphOptions.cpp -I. -Iinclude -std=c++14 -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 -DARM_COMPUTE_CL 254 #64 bit: 255 aarch64-linux-android-clang++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp utils/CommonGraphOptions.cpp -I. -Iinclude -std=c++14 -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 -DARM_COMPUTE_CL 256 257@note Due to some issues in older versions of the Arm® Mali™ OpenCL DDK (<= r13p0), we recommend to link arm_compute statically on Android. 258@note When linked statically the arm_compute_graph library currently needs the --whole-archive linker flag in order to work properly 259 260Then you need to do is upload the executable and the shared library to the device using ADB: 261 262 adb push neon_cnn_arm /data/local/tmp/ 263 adb push cl_sgemm_arm /data/local/tmp/ 264 adb push gc_absdiff_arm /data/local/tmp/ 265 adb shell chmod 777 -R /data/local/tmp/ 266 267And finally to run the example: 268 269 adb shell /data/local/tmp/neon_cnn_arm 270 adb shell /data/local/tmp/cl_sgemm_arm 271 adb shell /data/local/tmp/gc_absdiff_arm 272 273For 64bit: 274 275 adb push neon_cnn_aarch64 /data/local/tmp/ 276 adb push cl_sgemm_aarch64 /data/local/tmp/ 277 adb push gc_absdiff_aarch64 /data/local/tmp/ 278 adb shell chmod 777 -R /data/local/tmp/ 279 280And finally to run the example: 281 282 adb shell /data/local/tmp/neon_cnn_aarch64 283 adb shell /data/local/tmp/cl_sgemm_aarch64 284 adb shell /data/local/tmp/gc_absdiff_aarch64 285 286@note Examples accept different types of arguments, to find out what they are run the example with \a --help as an argument. If no arguments are specified then random values will be used to execute the graph. 287 288For example: 289 adb shell /data/local/tmp/graph_lenet --help 290 291In this case the first argument of LeNet (like all the graph examples) is the target (i.e 0 to run on Neon™, 1 to run on OpenCL if available, 2 to run on OpenCL using the CLTuner), the second argument is the path to the folder containing the npy files for the weights and finally the third argument is the number of batches to run. 292 293@section S1_4_macos Building for macOS 294 295The library was successfully natively built for Apple Silicon under macOS 11.1 using clang v12.0.0. 296 297To natively compile the library with accelerated CPU support: 298 299 scons Werror=1 -j8 neon=1 opencl=0 os=macos arch=armv8a build=native 300 301@note Initial support disables feature discovery through HWCAPS and thread scheduling affinity controls 302 303@section S1_5_bare_metal Building for bare metal 304 305For bare metal, the library was successfully built using linaro's latest (gcc-linaro-6.3.1-2017.05) bare metal toolchains: 306 - arm-eabi for armv7a 307 - aarch64-elf for armv8a 308 309Download 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/">armv8a</a>. 310 311@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 312 313@subsection S1_5_1_library How to build the library ? 314 315To cross-compile the library with Arm® Neon™ support for baremetal armv8a: 316 317 scons Werror=1 -j8 debug=0 neon=1 opencl=0 os=bare_metal arch=armv8a build=cross_compile cppthreads=0 openmp=0 standalone=1 318 319@subsection S1_5_2_examples How to manually build the examples ? 320 321Examples 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>. 322 323@section S1_6_windows_host Building on a Windows host system (cross-compile) 324 325Using `scons` directly from the Windows command line is known to cause 326problems. The reason seems to be that if `scons` is setup for cross-compilation 327it gets confused about Windows style paths (using backslashes). Thus it is 328recommended to follow one of the options outlined below. 329 330@subsection S1_6_1_ubuntu_on_windows Bash on Ubuntu on Windows (cross-compile) 331 332The best and easiest option is to use 333<a href="https://msdn.microsoft.com/en-gb/commandline/wsl/about">Ubuntu on Windows</a>. 334This feature is still marked as *beta* and thus might not be available. 335However, if it is building the library is as simple as opening a *Bash on 336Ubuntu on Windows* shell and following the general guidelines given above. 337 338@subsection S1_6_2_cygwin Cygwin (cross-compile) 339 340If the Windows subsystem for Linux is not available <a href="https://www.cygwin.com/">Cygwin</a> 341can be used to install and run `scons`, the minimum Cygwin version must be 3.0.7 or later. In addition 342to the default packages installed by Cygwin `scons` has to be selected in the installer. (`git` might 343also be useful but is not strictly required if you already have got the source 344code of the library.) Linaro provides pre-built versions of 345<a href="http://releases.linaro.org/components/toolchain/binaries/">GCC cross-compilers</a> 346that can be used from the Cygwin terminal. When building for Android the 347compiler is included in the Android standalone toolchain. After everything has 348been set up in the Cygwin terminal the general guide on building the library 349can be followed. 350 351@subsection S1_6_3_WoA Windows on ARM (native build) 352 353 Native builds on Windows are experimental and some features from the library interacting with the OS are missing. 354 355It's possible to build Compute Library natively on a windows system running on ARM. 356 357Windows on ARM(WoA) systems provide compatibility emulating x86 binaries on aarch64. Unfortunately Visual Studio 2022 does not work on aarch64 systems because it's an x86_64bit application and these binaries cannot be exectuted on WoA yet. 358 359Because we cannot use Visual Studio to build Compute Library we have to set up a native standalone toolchain to compile C++ code for arm64 on Windows. 360 361Native arm64 toolchain installation for WoA: 362- LLVM+Clang-12 which can be downloaded from: https://github.com/llvm/llvm-project/releases/download/llvmorg-12.0.0/LLVM-12.0.0-woa64.exe 363- Arm64 VC Runtime which can be downloaded from https://aka.ms/vs/17/release/vc_redist.arm64.exe 364 365- While full VS22 cannot be installed on WoA, we can install some components 366 -# Desktop development with C++ and all Arm64 components for Visual Studio, refer to: https://developer.arm.com/documentation/102528/0100/Install-Visual-Studio 367 -# VS22 build tools: https://visualstudio.microsoft.com/downloads/#build-tools-for-visual-studio-2022 368 369There are some additional tools we need to install to build Compute Library: 370 371- git https://git-scm.com/download/win 372- python 3 https://www.python.org/downloads/windows/ 373- scons can be installed with pip install scons 374 375In order to use clang to build windows binaries natively we have to initialize the environment variables from VS22 correctly so that the compiler could find the arm64 C++ libraries. This can be done by pressing the key windows + r and running the command: 376 377 cmd /k "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvarsx86_arm64.bat" 378 379To build Compute Library type: 380 381 scons opencl=0 neon=1 os=windows examples=0 validation_tests=1 benchmark_examples=0 build=native arch=armv8a Werror=0 exceptions=1 standalone=1 382 383@section S1_7_cl_requirements OpenCL DDK Requirements 384 385@subsection S1_7_1_cl_hard_requirements Hard Requirements 386 387Compute Library requires OpenCL 1.1 and above with support of non uniform workgroup sizes, which is officially supported in the Arm® Mali™ OpenCL DDK r8p0 and above as an extension (respective extension flag is \a -cl-arm-non-uniform-work-group-size). 388 389Enabling 16-bit floating point calculations require \a cl_khr_fp16 extension to be supported. All Arm® Mali™ GPUs with compute capabilities have native support for half precision floating points. 390 391@subsection S1_7_2_cl_performance_requirements Performance improvements 392 393Integer dot product built-in function extensions (and therefore optimized kernels) are available with Arm® Mali™ OpenCL DDK r22p0 and above for the following GPUs : G71, G76. The relevant extensions are \a cl_arm_integer_dot_product_int8, \a cl_arm_integer_dot_product_accumulate_int8 and \a cl_arm_integer_dot_product_accumulate_int16. 394 395OpenCL kernel level debugging can be simplified with the use of printf, this requires the \a cl_arm_printf extension to be supported. 396 397SVM allocations are supported for all the underlying allocations in Compute Library. To enable this OpenCL 2.0 and above is a requirement. 398 399@section S1_8_experimental_builds Experimental Bazel and CMake builds 400 401In addition to the scons build the repository includes experimental Bazel and CMake builds. 402Both are similar to the scons multi_isa build. It compiles all libraries with Neon (TM) support, as well as SVE and SVE2 libraries. The build is CPU only, not including OpenCL support. Both were successfully built with gcc / g++ version 10.2. 403 404@subsection S1_8_1_bazel_build Bazel build 405 406@subsubsection S1_8_1_1_file_structure File structure 407 408File structure for all files included in the Bazel build: 409 410 . 411 ├── .bazelrc 412 ├── BUILD 413 ├── WORKSPACE 414 ├── arm_compute 415 │ └── BUILD 416 ├── examples 417 │ └── BUILD 418 ├── include 419 │ └── BUILD 420 ├── scripts 421 │ ├── print_version_file.py 422 │ └── BUILD 423 ├── src 424 │ └── BUILD 425 ├── support 426 │ └── BUILD 427 ├── tests 428 │ ├── BUILD 429 │ └── framework 430 │ └── BUILD 431 └── utils 432 └── BUILD 433 434@subsubsection S1_8_1_2_build_options Build options 435 436Available build options: 437 438 - debug: Enable ['-O0','-g','-gdwarf-2'] compilation flags 439 - Werror: Enable -Werror compilation flag 440 - logging: Enable logging 441 - cppthreads: Enable C++11 threads backend 442 - openmp: Enable OpenMP backend 443 444@subsubsection S1_8_1_3_example_builds Example builds 445 446Build everything (libraries, examples, tests): 447 448 bazel build //... 449 450Build libraries: 451 452 bazel build //:all 453 454Build arm_compute only: 455 456 bazel build //:arm_compute 457 458Build examples: 459 460 bazel build //examples:all 461 462Build resnet50 example: 463 464 bazel build //examples:graph_resnet50 465 466Build validation and benchmarking: 467 468 bazel build //tests:all 469 470@subsection S1_8_2_cmake_build CMake build 471 472@subsubsection S1_8_2_1_file_structure File structure 473 474File structure for all files included in the CMake build: 475 476 . 477 ├── CMakeLists.txt 478 ├── cmake 479 │ ├── Options.cmake 480 │ ├── Version.cmake 481 │ └── toolchains 482 │ └── aarch64_linux_toolchain.cmake 483 ├── examples 484 │ └── CMakeLists.txt 485 ├── src 486 │ └── CMakeLists.txt 487 └── tests 488 ├── CMakeLists.txt 489 ├── benchmark 490 │ └── CMakeLists.txt 491 └── validation 492 └── CMakeLists.txt 493 494@subsubsection S1_8_2_2_build_options Build options 495 496Available build options: 497 498 - DEBUG: Enable ['-O0','-g','-gdwarf-2'] compilation flags 499 - WERROR: Enable -Werror compilation flag 500 - EXCEPTIONS: If disabled ARM_COMPUTE_EXCEPTIONS_DISABLED is enabled 501 - LOGGING: Enable logging 502 - BUILD_EXAMPLES: Build examples 503 - BUILD_TESTING: Build tests 504 - CPPTHREADS: Enable C++11 threads backend 505 - OPENMP: Enable OpenMP backend 506 507@subsubsection S1_8_2_3_example_builds Example builds 508 509To build libraries, examples and tests: 510 511 mkdir build 512 cd build 513 cmake .. -DOPENMP=1 -DWERROR=0 -DDEBUG=0 -DBUILD_EXAMPLES=1 -DBUILD_TESTING=1 -DCMAKE_INSTALL_LIBDIR=. 514 cmake --build . -j32 515 516*/ 517} // namespace arm_compute 518