1 2> **⚠ Important** 3> From release 22.05: 'master' branch has been replaced with 'main' following our inclusive language update, more information [here](https://arm-software.github.io/ComputeLibrary/v23.02.1/contribution_guidelines.xhtml#S5_0_inc_lang). 4 5> **⚠ Important** 6> From release 22.08: armv7a with Android build will no longer be tested or maintained. 7 8> **⚠ Important** 9> From release 23.02: The 23.02 release introduces a change to the default tensor extend padding behavior. 10> To remain compatible with previous behavior, users will need to set the new flag `ITensorInfo::lock_paddings()` on 11> tensors for which paddings should not be extended, such as the input and output of the model that need to be mapped to 12> a camera frame or frame buffer. 13 14<br> 15<div align="center"> 16 <img src="https://raw.githubusercontent.com/ARM-software/ComputeLibrary/gh-pages/ACL_logo.png"/><br><br> 17</div> 18 19# Compute Library  20 21 22The Compute Library is a collection of low-level machine learning functions optimized for Arm® Cortex®-A, Arm® Neoverse® and Arm® Mali™ GPUs architectures.<br> 23 24The library provides superior performance to other open source alternatives and immediate support for new Arm® technologies e.g. SVE2. 25 26Key Features: 27 28- Open source software available under a permissive MIT license 29- Over 100 machine learning functions for CPU and GPU 30- Multiple convolution algorithms (GeMM, Winograd, FFT, Direct and indirect-GeMM) 31- Support for multiple data types: FP32, FP16, INT8, UINT8, BFLOAT16 32- Micro-architecture optimization for key ML primitives 33- Highly configurable build options enabling lightweight binaries 34- Advanced optimization techniques such as kernel fusion, Fast math enablement and texture utilization 35- Device and workload specific tuning using OpenCL tuner and GeMM optimized heuristics 36 37<br> 38 39| Repository | Link | 40| ----------- | ---------------------------------------------------------------- | 41| Release | https://github.com/arm-software/ComputeLibrary | 42| Development | https://review.mlplatform.org/#/admin/projects/ml/ComputeLibrary | 43 44<br> 45 46## Documentation 47[](https://arm-software.github.io/ComputeLibrary/v23.02.1) 48 49> Note: The documentation includes the reference API, changelogs, build guide, contribution guide, errata, etc. 50 51<br> 52 53## Pre-built binaries 54All the binaries can be downloaded from [here](https://github.com/ARM-software/ComputeLibrary/releases) or from the tables below. 55 56<br> 57 58| Platform | Operating System | Release archive (Download) | 59| -------------- | ---------------- | -------------------------- | 60| Raspberry Pi 4 | Linux 32bit | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-armv7a-neon.tar.gz) | 61| Raspberry Pi 4 | Linux 64bit | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-neon.tar.gz) | 62| Odroid N2 | Linux 64bit | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-neon.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-cl.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-neon-cl.tar.gz) | 63| HiKey960 | Linux 64bit | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-neon.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-cl.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-neon-cl.tar.gz) | 64 65<br> 66 67| Architecture | Operating System | Release archive (Download) | 68| ------------ | ---------------- | -------------------------- | 69| armv7 | Linux | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-armv7a-neon.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-armv7a-cl.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-armv7a-neon-cl.tar.gz) | 70| arm64-v8a | Android | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-android-arm64-v8a-neon.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-android-arm64-v8a-cl.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-android-arm64-v8a-neon-cl.tar.gz) | 71| arm64-v8a | Linux | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-neon.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-cl.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8a-neon-cl.tar.gz) | 72| arm64-v8.2-a | Android | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-android-arm64-v8.2-a-neon.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-android-arm64-v8.2-a-cl.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-android-arm64-v8.2-a-neon-cl.tar.gz) | 73| arm64-v8.2-a | Linux | [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8.2-a-neon.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8.2-a-cl.tar.gz) [](https://github.com/ARM-software/ComputeLibrary/releases/download/v23.02.1/arm_compute-v23.02.1-bin-linux-arm64-v8.2-a-neon-cl.tar.gz) | 74 75<br> 76 77Please refer to the following link for more pre-built binaries: [](https://github.com/ARM-software/ComputeLibrary/releases/tag/v23.02.1) 78 79Pre-build binaries are generated with the following security / good coding practices related flags: 80> -Wall, -Wextra, -Wformat=2, -Winit-self, -Wstrict-overflow=2, -Wswitch-default, -Woverloaded-virtual, -Wformat-security, -Wctor-dtor-privacy, -Wsign-promo, -Weffc++, -pedantic, -fstack-protector-strong 81 82## Supported Architectures/Technologies 83 84- Arm® CPUs: 85 - Arm® Cortex®-A processor family using Arm® Neon™ technology 86 - Arm® Neoverse® processor family 87 - Arm® Cortex®-R processor family with Armv8-R AArch64 architecture using Arm® Neon™ technology 88 - Arm® Cortex®-X1 processor using Arm® Neon™ technology 89 90- Arm® Mali™ GPUs: 91 - Arm® Mali™-G processor family 92 - Arm® Mali™-T processor family 93 94- x86 95 96<br> 97 98## Supported Systems 99 100- Android™ 101- Bare Metal 102- Linux® 103- OpenBSD® 104- macOS® 105- Tizen™ 106 107<br> 108 109## Resources 110- [Tutorial: Running AlexNet on Raspberry Pi with Compute Library](https://community.arm.com/processors/b/blog/posts/running-alexnet-on-raspberry-pi-with-compute-library) 111- [Gian Marco's talk on Performance Analysis for Optimizing Embedded Deep Learning Inference Software](https://www.embedded-vision.com/platinum-members/arm/embedded-vision-training/videos/pages/may-2019-embedded-vision-summit) 112- [Gian Marco's talk on optimizing CNNs with Winograd algorithms at the EVS](https://www.embedded-vision.com/platinum-members/arm/embedded-vision-training/videos/pages/may-2018-embedded-vision-summit-iodice) 113- [Gian Marco's talk on using SGEMM and FFTs to Accelerate Deep Learning](https://www.embedded-vision.com/platinum-members/arm/embedded-vision-training/videos/pages/may-2016-embedded-vision-summit-iodice) 114 115<br> 116 117## Experimental builds 118 119**⚠ Important** Bazel and CMake builds are experimental CPU only builds, please see the [documentation](https://arm-software.github.io/ComputeLibrary/v23.02.1/how_to_build.xhtml) for more details. 120 121<br> 122 123## How to contribute 124 125Contributions to the Compute Library are more than welcome. If you are interested on contributing, please have a look at our [how to contribute guidelines](https://arm-software.github.io/ComputeLibrary/v23.02.1/contribution_guidelines.xhtml). 126 127### Developer Certificate of Origin (DCO) 128Before the Compute Library accepts your contribution, you need to certify its origin and give us your permission. To manage this process we use the Developer Certificate of Origin (DCO) V1.1 (https://developercertificate.org/) 129 130To indicate that you agree to the the terms of the DCO, you "sign off" your contribution by adding a line with your name and e-mail address to every git commit message: 131 132```Signed-off-by: John Doe <john.doe@example.org>``` 133 134You must use your real name, no pseudonyms or anonymous contributions are accepted. 135 136### Public mailing list 137For technical discussion, the ComputeLibrary project has a public mailing list: acl-dev@lists.linaro.org 138The list is open to anyone inside or outside of Arm to self subscribe. In order to subscribe, please visit the following website: 139https://lists.linaro.org/mailman3/lists/acl-dev.lists.linaro.org/ 140 141<br> 142 143## License and Contributions 144 145The software is provided under MIT license. Contributions to this project are accepted under the same license. 146 147### Other Projects 148This project contains code from other projects as listed below. The original license text is included in those source files. 149 150* The OpenCL header library is licensed under Apache License, Version 2.0, which is a permissive license compatible with MIT license. 151 152* The half library is licensed under MIT license. 153 154* The libnpy library is licensed under MIT license. 155 156* The stb image library is either licensed under MIT license or is in Public Domain. It is used by this project under the terms of MIT license. 157 158<br> 159 160## Trademarks and Copyrights 161 162Android is a trademark of Google LLC. 163 164Arm, Cortex, Mali and Neon are registered trademarks or trademarks of Arm Limited (or its subsidiaries) in the US and/or elsewhere. 165 166Bazel is a trademark of Google LLC., registered in the U.S. and other 167countries. 168 169CMake is a trademark of Kitware, Inc., registered in the U.S. and other 170countries. 171 172Linux® is the registered trademark of Linus Torvalds in the U.S. and other countries. 173 174Mac and macOS are trademarks of Apple Inc., registered in the U.S. and other 175countries. 176 177Tizen is a registered trademark of The Linux Foundation. 178 179
