diff --git a/CHANGELOG.md b/CHANGELOG.md index bd973ec..b7e7bee 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -18,9 +18,11 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/). ## [0.28.1.5] - YYYY-MM-DD -### Fixed +### Changed -- Fixed ReduceScatter errors caused by OneCCL api change. +- Upgraded Driver version and OneAPI version. +- Turned on TensorFlow NextPluggableDevice by default. +- Fixed inplace reduce_scatter for OneCCL api change. ## [0.28.1.4] - YYYY-MM-DD diff --git a/README.md b/README.md index 4644d2e..24c8179 100644 --- a/README.md +++ b/README.md @@ -3,14 +3,14 @@ Intel® Optimization for Horovod* is the distributed training framework for Tens ## Install ### Hardware Requirements - - Intel® Data Center GPU Max Series, Driver Version: [803](https://dgpu-docs.intel.com/releases/LTS_803.29_20240131.html) + - Intel® Data Center GPU Max Series, Driver Version: [803](https://dgpu-docs.intel.com/releases/LTS_803.63_20240617.html) |Software|Installation requirement| |-|-| |Intel® oneAPI Base Toolkit|[Install Intel® oneAPI Base Toolkit](https://github.com/intel/intel-extension-for-tensorflow/tree/r2.15/docs/install/install_for_xpu.md#install-oneapi-base-toolkit-packages)| -|TensorFlow|[Install tensorflow 2.15.0](https://www.tensorflow.org/install)| +|TensorFlow|[Install tensorflow 2.15.1](https://www.tensorflow.org/install)| |Intel® Extension for TensorFlow*|[Install Intel® Extension for TensorFlow*](https://github.com/intel/intel-extension-for-tensorflow/tree/r2.15#install) | |System|Ubuntu 22.04, SUSE Linux Enterprise Server(SLES) 15 SP3/SP4| |Python|3.9-3.11| @@ -20,7 +20,7 @@ Intel® Optimization for Horovod* is the distributed training framework for Tens |OS|Intel GPU|Install Intel GPU Driver| |-|-|-| -|Ubuntu 22.04, RedHat 8.6, SLES 15 SP3/SP4|Intel® Data Center GPU Max Series| Refer to the [Installation Guides](https://dgpu-docs.intel.com/installation-guides/index.html#intel-data-center-gpu-max-series) for latest driver installation. If install the verified Intel® Data Center GPU Max Series/Intel® Data Center GPU Flex Series [803](https://dgpu-docs.intel.com/releases/LTS_803.29_20240131.html), please append the specific version after components.| +|Ubuntu 22.04, RedHat 8.6, SLES 15 SP3/SP4|Intel® Data Center GPU Max Series| Refer to the [Installation Guides](https://dgpu-docs.intel.com/installation-guides/index.html#intel-data-center-gpu-max-series) for latest driver installation. If install the verified Intel® Data Center GPU Max Series/Intel® Data Center GPU Flex Series [803](https://dgpu-docs.intel.com/releases/LTS_803.63_20240617.html), please append the specific version after components.| ### Installation Channel: @@ -36,7 +36,7 @@ Intel® Optimization for Horovod* can be installed through the following channel Installing Intel® Optimization for Horovod* with different frameworks is feasible. You could choose Intel® Extension for TensorFlow* as dependency. 1. Installing Intel® Extension for TensorFlow* and Intel® Optimization for Horovod* with command:
```bash - pip install tensorflow==2.15.0 + pip install tensorflow==2.15.1 pip install --upgrade intel-extension-for-tensorflow[xpu] pip install intel-optimization-for-horovod ``` diff --git a/docs/intel_gpus.rst b/docs/intel_gpus.rst index cbe6958..43870d4 100644 --- a/docs/intel_gpus.rst +++ b/docs/intel_gpus.rst @@ -18,7 +18,7 @@ Install Intel GPU driver Refer to `Installation Guides `__ for latest driver installation. -Recommend to install verified driver version `803 `__ for hardware platforms: +Recommend to install verified driver version `803 `__ for hardware platforms: Intel® Data Center GPU Max Series and Intel® Data Center GPU Flex Series 170. @@ -33,8 +33,8 @@ There are necessary 3 components for horovod on Intel GPU: .. code-block:: bash - $ wget https://registrationcenter-download.intel.com/akdlm/IRC_NAS/fdc7a2bc-b7a8-47eb-8876-de6201297144/l_BaseKit_p_2024.1.0.596.sh - $ sudo sh ./l_BaseKit_p_2024.1.0.596.sh + $ wget https://registrationcenter-download.intel.com/akdlm/IRC_NAS/e6ff8e9c-ee28-47fb-abd7-5c524c983e1c/l_BaseKit_p_2024.2.1.100_offline.sh + $ sudo sh l_BaseKit_p_2024.2.1.100_offline.sh For any more details, follow the `procedures `__. @@ -51,7 +51,7 @@ Install Deep Learning Frameworks To bring Intel GPU devices into **TensorFlow** community for AI workload acceleration, Intel promoted a new user visible ``XPU`` device type as a device abstraction for Intel computation architectures and implemented corresponding device runtime in **Intel® Extension for TensorFlow**. -To use **TensorFlow** and **Intel® Extension for TensorFlow**, please follow `Tensorflow 2.15.0 installation `__ and `Intel® Extension for TensorFlow* 2.15 installation `__. +To use **TensorFlow** and **Intel® Extension for TensorFlow**, please follow `Tensorflow 2.15.1 installation `__ and `Intel® Extension for TensorFlow* 2.15 installation `__. Install the Horovod pip package ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ diff --git a/xpu_docs/how_to_build.md b/xpu_docs/how_to_build.md index d4c54c2..68be33e 100644 --- a/xpu_docs/how_to_build.md +++ b/xpu_docs/how_to_build.md @@ -20,9 +20,9 @@ Need to install components of Intel® oneAPI Base Toolkit: - Intel® oneAPI MPI Library (IntelMPI) ```bash -$ wget https://registrationcenter-download.intel.com/akdlm/IRC_NAS/fdc7a2bc-b7a8-47eb-8876-de6201297144/l_BaseKit_p_2024.1.0.596.sh +$ wget https://registrationcenter-download.intel.com/akdlm/IRC_NAS/e6ff8e9c-ee28-47fb-abd7-5c524c983e1c/l_BaseKit_p_2024.2.1.100_offline.sh # 3 components are necessary: DPC++/C++ Compiler with DPC++ Libiary, oneMKL and oneCCL(IntelMPI will be installed automatically as oneCCL's dependency). -$ sudo sh ./l_BaseKit_p_2024.1.0.596.sh +$ sudo sh l_BaseKit_p_2024.2.1.100_offline.sh ``` For any more details, please follow the procedure in https://www.intel.com/content/www/us/en/developer/tools/oneapi/base-toolkit.html. @@ -44,7 +44,7 @@ source /path to basekit/intel/oneapi/ccl/latest/env/vars.sh Intel® Optimization for Horovod* depends on TensorFlow* to build from source. ```bash -$ pip install tensorflow==2.15.0 +$ pip install tensorflow==2.15.1 ```