LASR Procedure
- Syntax
- Overview
- Examples Start a ServerStarting a Server with Logging OptionsUsing the SAS Data in HDFS EngineLoad a Table from Teradata to MemoryLoad a Table from Greenplum to MemoryUnload a Table from MemoryStopping a ServerWorking with User-Defined FormatsWorking with User-Defined Formats and the FMTLIBXML= OptionSaving a Table to HDFS
PERFORMANCE Statement
The PERFORMANCE statement defines performance parameters for multithreaded and distributed computing.
| Examples: | Load a Table from Teradata to Memory |
Syntax
Performance Statement Options
COMMIT=
specifies that periodic updates are written to the SAS log when observations are sent from the client to the server instance. Whenever the number of observations sent exceeds an integer multiple of the COMMIT= size, a message is written to the SAS log. The message indicates the actual number of observations distributed and not an integer multiple of the COMMIT= size.
DATASERVER=
specifies the host to use for a database connection. This option is used in Teradata deployments so that the LASR procedure compares this host name with the host name that is specified in the SERVER= option in the LIBNAME statement. If you do not specify the DATASERVER= option, the host to use for the database connection is determined from the GRIDDATASERVER= environment variable.
HOST=
specifies the grid host to use for the server instance. Enclose the host name in quotation marks. If you do not specify the HOST= option, it is determined from the GRIDHOST= environment variable.
| Alias | GRIDHOST= |
INSTALL=
specifies the path to the TKGrid software on the grid host. If you do not specify this option, it is determined from the GRIDINSTALLLOC= environment variable.
| Alias | INSTALLOC= |
LASRSERVER=
specifies the server to use. Provide the fully qualified path to the server description file.
| Alias | LASR= |
MODE= SYM | ASYM
specifies whether the server runs in symmetric (SYM) mode or asymmetric (ASYM) mode. Asymmetric mode is useful when you want to read data in parallel from a massively parallel processing (MPP) data appliance that is not co-located with the server.
| Alias | GRIDMODE= |
| Default | SYM |
NODES=
specifies the number of machines in the cluster to use for the server instance. Specify ALL to calculate the number automatically.
| Alias | NNODES= |
| Restriction | This option has no effect when you use a third-party vendor database as a co-located data provider and you specify the CREATE= and DATA= options in the PROC LASR statement. When you use a third-party vendor database as a co-located data provider, you must use all of the machines to read data from the database. |
NTHREADS=
specifies the number of threads for analytic computations and overrides the SAS system option THREADS | NOTHREADS. By default, the server uses one thread for each CPU core that is available on each machine in the cluster. Use this option to throttle the number of CPU cores that are used on each machine.
TIMEOUT=
specifies the time in seconds for the LASR procedure to wait for a connection to the grid host and establish a connection back to the client. The default value is 120 seconds. If jobs are submitted through workload management tools that might suspend access to the grid host for a longer period, you might want to increase the value.