SAP SAP Data Intelligence 3 on Rancher Kubernetes Engine 2 using VMware vSAN and vSphere SUSE Linux Enterprise Server 15 SP4 Rancher Kubernetes Engine 2 SAP Data Intelligence 3 Dr. Ulrich Schairer, (SUSE) 1 SAP Data Intelligence 3 on Rancher Kubernetes Engine 2 using VMware vSAN and vSphere SAP Data Intelligence 3 on Rancher Kubernetes Engine 2 using VMware vSAN and vSphere Date: 2023-07-24 SAP possi- ble errors or the consequences thereof. 2 SAP Data Intelligence 3 on Rancher Kubernetes Engine 2 using VMware vSAN and vSphere Contents 1 Introduction 4 2 Requirements 5 3 Preparations 7

从FM到DeepFM rt 增加了10 倍怎么优化？ 深度学习应用主要的挑战： 2.模型效果优 化困难 1.方案复杂  训练优化:  数据并行  模型并行  推理优化: Blade  推荐模型优化: 千亿特征 3. 工程优化 RingAllReduce + 层级级联 EasyVision 多机多卡性能对比 工程优化: 数据并行  M6模型  Transformer模型: RapidFormer性能 工程优化: 模型并行(Whale)  FP16 / Int8  模型剪枝  Op融合(Fusion Stitch)  MILR: Blade Disc 工程优化: Blade模型推理 Dynamic Shape Compiler for Machine Learning Workloads EmbeddingVariable [No Hash 病预测等） Infrastructure PAI平台（Platform of Artificial Intelligence） • 一键部署、弹性扩缩 • 多框架、多语言 • 推理优化Blade • 多维度监控+报警 • 自定义镜像 • 全托管+半托管 • 分布式训练优化 • 超大资源池 智能标注 可视化建模(Designer) 分布式训练(DLC) 在线服务(EAS)

does not need Docker containers when used with distributions such as K3s and Rancher Kubernetes Engine 2 (RKE2). For installations that want an even smaller attack surface, SUSE Rancher can utilize an the hosted hyperscaler Kubernetes services. With RHACM, operators also get access to a policy engine via GitOps help manage clusters at scale. OpenShift clusters will also have full monitoring capabilities Additionally, SUSE Rancher-managed Amazon EKS, Microsoft AKS and Google GKE deployments support templating and CIS benchmark scanning to maintain high security and minimize configuration drift between

attestation is desired, pilot-agent template generation should be revisited to ensure that all templating performs output encoding in a context-appropriate manner. Sidecar isolation is an important boundary

sion Notes SciPy 1.7.1 Miscellaneous statistical functions numba 0.53.1 Alternative execution engine for rolling operations (see Enhancing Perfor- mance) xarray 0.19.0 pandas-like API for N-dimensional 4.16 SQL support for databases other than sqlite psycopg2 2.8.6 PostgreSQL engine for sqlalchemy pymysql 1.0.2 MySQL engine for sqlalchemy Other data sources Dependency Minimum Version Notes PyTables ase.py:3794, in Index.get_loc(self, key, method,␣ ˓→tolerance) 3793 try: -> 3794 return self._engine.get_loc(casted_key) 3795 except KeyError as err: File /pandas/pandas/_libs/index.pyx:138, in pandas

users should pay close attention to. 1.4.1 File parsing new features The delimited file parsing engine (the guts of read_csv and read_table) has been rewritten from the ground up and now uses a fraction both simple axis indexing and multi-level / hierarchical axis indexing • An integrated group by engine for aggregating and transforming data sets • Date range generation (date_range) and custom date can extract some data into a DataFrame. In the following example, we use the SQlite SQL database engine. You can use a temporary SQLite database where data are stored in “memory”. Just do: import sqlite3

users should pay close attention to. 1.6.1 File parsing new features The delimited file parsing engine (the guts of read_csv and read_table) has been rewritten from the ground up and now uses a fraction both simple axis indexing and multi-level / hierarchical axis indexing • An integrated group by engine for aggregating and transforming data sets • Date range generation (date_range) and custom date default Excel writer engine for ’xls’ files. Available options: ’xlwt’ (the default). io.excel.xlsm.writer: [default: openpyxl] [currently: openpyxl] : string The default Excel writer engine for ’xlsm’ files

specified with delim_whitespace=True in read_csv()/read_table() (GH6607) • Raise ValueError when engine=’c’ specified with unsupported options in read_csv()/read_table() (GH6607) • Raise ValueError when a SQLAlchemy engine to the sql functions. To connect with SQLAlchemy you use the create_engine() function to create an engine object from database URI. You only need to create the engine once per database Release 0.14.0 In [43]: from sqlalchemy import create_engine # Create your connection. In [44]: engine = create_engine(’sqlite:///:memory:’) This engine can then be used to write or read data to/from this

sion Notes SciPy 1.4.1 Miscellaneous statistical functions numba 0.50.1 Alternative execution engine for rolling operations (see Enhancing Perfor- mance) xarray 0.15.1 pandas-like API for N-dimensional 4.0 SQL support for databases other than sqlite psycopg2 2.8.4 PostgreSQL engine for sqlalchemy pymysql 0.10.1 MySQL engine for sqlalchemy Other data sources Dependency Minimum Version Notes PyTables \t for read_table()] Delimiter to use. If sep is None, the C engine cannot automatically detect the separator, but the Python parsing engine can, meaning the latter will be used and automatically detect

default Text type for string columns: from sqlalchemy.types import String data.to_sql(’data_dtype’, engine, dtype={’Col_1’: String}) • Series.all and Series.any now support the level and skipna parameters read_sql_table and to_sql (GH7441, GH7952). For example: df.to_sql(’table’, engine, schema=’other_schema’) pd.read_sql_table(’table’, engine, schema=’other_schema’) • Added support for writing NaN values with now has a keyword parameter float_precision which specifies which floating-point converter the C engine should use during parsing, see here (GH8002, GH8044) • Added searchsorted method to Series objects