AMP will select an optimal set of operations to cast to FP16. FP16 operations require 2X reduced memory bandwidth (resulting in a 2X speedup for bandwidth-bound operations like most pointwise ops) and 2X AMP will select an optimal set of operations to cast to FP16. FP16 operations require 2X reduced memory bandwidth (resulting in a 2X speedup for bandwidth-bound operations like most pointwise ops) and 2X full iteration CUDA graph capture including gradient AllReduce, Optimizer, and Parameter AllGather operations could fail with a CUDA error. We recommend reducing the scope of the CUDA graph capture as a workaround

live stream into batches of X seconds. • Treats each batch as RDDs and processes them using RDD operations. • Discretized Stream Processing (DStream) 7 / 79 Spark Streaming ▶ Run a streaming computation live stream into batches of X seconds. • Treats each batch as RDDs and processes them using RDD operations. • Discretized Stream Processing (DStream) 7 / 79 Spark Streaming ▶ Run a streaming computation live stream into batches of X seconds. • Treats each batch as RDDs and processes them using RDD operations. • Discretized Stream Processing (DStream) 7 / 79 DStream (1/2) ▶ DStream: sequence of RDDs

users to execute Python native functions. See also the latest User- defined Functions and Row-based Operations. The first example is UDFs used in Table API & SQL [20]: from pyflink.table.udf import udf # sql_query("SELECT plus_one(id) FROM {}".format(table)).to_pandas() Another example is UDFs used in Row-based Operations [23]: from pyflink.common.types import Row @udf(result_type=DataTypes.ROW([DataTypes.FIELD("id" QueryOperationConverter$SingleRelVisitor. ˓→visit(QueryOperationConverter.java:154) at org.apache.flink.table.operations.CatalogQueryOperation. ˓→accept(CatalogQueryOperation.java:68) at org.apache.flink.table.planner

users to execute Python native functions. See also the latest User- defined Functions and Row-based Operations. The first example is UDFs used in Table API & SQL [20]: from pyflink.table.udf import udf # sql_query("SELECT plus_one(id) FROM {}".format(table)).to_pandas() Another example is UDFs used in Row-based Operations [23]: from pyflink.common.types import Row @udf(result_type=DataTypes.ROW([DataTypes.FIELD("id" QueryOperationConverter$SingleRelVisitor. ˓→visit(QueryOperationConverter.java:154) at org.apache.flink.table.operations.CatalogQueryOperation. ˓→accept(CatalogQueryOperation.java:68) at org.apache.flink.table.planner

Formatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 1.3.2.12 Window Binary Corr/Cov operations return a MultiIndex DataFrame . . . . . . . . 31 1.3.2.13 HDFStore where string comparison . . . . . . . . . . . . . . . . . . . . . . . . . 74 1.6.2.8 Index + / - no longer used for set operations . . . . . . . . . . . . . . . . . . . . . 76 1.6.2.9 Index.difference and .symmetric_difference . . . . . . . . . . . . . . . . . . . 93 1.7.1.2 .groupby(..) syntax with window and resample operations . . . . . . . . . . . 93 1.7.1.3 Method chaininng improvements . . . . . . . . . . . . . . .

Formatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 1.2.2.12 Window Binary Corr/Cov operations return a MultiIndex DataFrame . . . . . . . . 29 1.2.2.13 HDFStore where string comparison . . . . . . . . . . . . . . . . . . . . . . . . . 72 1.5.2.8 Index + / - no longer used for set operations . . . . . . . . . . . . . . . . . . . . . 74 1.5.2.9 Index.difference and .symmetric_difference . . . . . . . . . . . . . . . . . . . 91 1.6.1.2 .groupby(..) syntax with window and resample operations . . . . . . . . . . . 91 1.6.1.3 Method chaininng improvements . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 2.1.5 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 2 . . . . . . . . 199 2.3.3 Accelerated operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 2.3.4 Flexible binary operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 651 2.12.7 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 653 2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 2.1.5 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 2 . . . . . . . . 192 2.3.3 Accelerated operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 2.3.4 Flexible binary operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 622 2.12.7 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625 2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 2.1.5 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 2 . . . . . . . . 199 2.3.3 Accelerated operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 2.3.4 Flexible binary operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 650 2.12.7 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 652 2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 5.5 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 8.3 Accelerated operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 8.4 Flexible binary operations . . . . . . . . . . . . . . . . . you in computations • Powerful, flexible group by functionality to perform split-apply-combine operations on data sets, for both ag- gregating and transforming data • Make it easy to convert ragged, differently-indexed