From 03ab06963badb64f494a1636a4c07238ea4be011 Mon Sep 17 00:00:00 2001
From: "C.J. Collier" <cjac@google.com>
Date: Tue, 24 Sep 2024 12:06:48 -0700
Subject: [PATCH] initial commit of ipex action + README

---
 ipex/README.md    | 33 +++++++++++++++++++++++++++++++++
 ipex/ipex-init.sh | 39 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 72 insertions(+)
 create mode 100644 ipex/README.md
 create mode 100644 ipex/ipex-init.sh

diff --git a/ipex/README.md b/ipex/README.md
new file mode 100644
index 000000000..b7a82ce5e
--- /dev/null
+++ b/ipex/README.md
@@ -0,0 +1,33 @@
+
+## In Intel's own words:
+
+Intel® Extension for PyTorch* extends PyTorch* with the latest
+performance optimizations for Intel hardware. Optimizations take
+advantage of Intel® Advanced Vector Extensions 512 (Intel® AVX-512)
+Vector Neural Network Instructions (VNNI) and Intel® Advanced Matrix
+Extensions (Intel® AMX) on Intel CPUs as well as Intel XeMatrix
+Extensions (XMX) AI engines on Intel discrete GPUs. Moreover, Intel®
+Extension for PyTorch* provides easy GPU acceleration for Intel
+discrete GPUs through the PyTorch* xpu device.
+
+The extension can be loaded as a Python module for Python programs or
+linked as a C++ library for C++ programs. In Python scripts, users can
+enable it dynamically by importing intel_extension_for_pytorch.
+
+## This action:
+
+The ipex initialization action installs the appropriate python
+libraries and installs the spark-tfrecord jar for the current scala
+runtime into the cluster's classpath.
+
+### Metadata arguments
+* metadata name: spark-tfrecord-version
+* default value: 0.4.0
+  Used to specify the version of the spark-tfrecord jar.  See url
+  https://central.sonatype.com/artifact/com.linkedin.sparktfrecord/spark-tfrecord_2.12/versions
+  https://central.sonatype.com/artifact/com.linkedin.sparktfrecord/spark-tfrecord_2.13/versions
+
+* metadata name: scala-version
+* defaut value: 2.12
+  Used to specify the version of scala.  See url
+  https://github.com/scala/scala/releases
diff --git a/ipex/ipex-init.sh b/ipex/ipex-init.sh
new file mode 100644
index 000000000..527cc4749
--- /dev/null
+++ b/ipex/ipex-init.sh
@@ -0,0 +1,39 @@
+#!/bin/bash
+
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS-IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+# This script installs ipex acceleration drivers and does not collects GPU utilization metrics at this time
+
+set -euxo pipefail
+
+mkdir -p /home/.cache
+chmod -R 777 /home/.cache
+
+pip install tiktoken
+pip install tldextract
+pip install zstandard
+pip install transformers
+pip install -U optimum[neural-compressor] intel-extension-for-transformers
+pip install intel_extension_for_pytorch
+pip install fasttext
+export CUDA_VISIBLE_DEVICES=""
+
+DEFAULT_SCALA_VER=2.12
+DEFAULT_TFRECORD_VER=0.4.0
+SCALA_VER=$(get_metadata_attribute 'scala-version' "${DEFAULT_SCALA_VER}")
+TFRECORD_VER=$(get_metadata_attribute 'spark-tfrecord-version' "${DEFAULT_TFRECORD_VER}")
+
+JAR_FILENAME="spark-tfrecord_${SCALA_VER}-${TFRECORD_VER}.jar"
+wget "https://repo1.maven.org/maven2/com/linkedin/sparktfrecord/spark-tfrecord_${SCALA_VER}/${TFRECORD_VER}/${JAR_FILENAME}"
+mv "${JAR_FILENAME}" /usr/lib/spark/jars/