HP MLIB for Itanium Linux Version 9.6.1 Release Note

Compatibility Information

HP MLIB for Itanium Linux Version 9.6.1 Release Note 23

The following command lines show the C-like shell syntax and Bourne-like shell syntax to

use when setting the variable to two processors:

For C-like shells:

% setenv MLIB_NUMBER_OF_THREADS 2

For Bourne-like shells:

% export MLIB_NUMBER_OF_THREADS=2

MLIB_NUMBER_OF_THREADS is examined on the first call to a parallelized VECLIB

subprogram to establish the default parallel action within VECLIB.

Use the subroutine MLIB_SETNUMTHREADS to override the environment variable

MLIB_NUMBER_OF_THREADS. Refer to the mlib_setnumthreads(3m) man page for usage

information.

2. Use the subroutine MLIB_SETNUMTHREADS.

You can call this subroutine at any time to set the maximum number of parallel threads

used in subsequent VECLIB or LAPACK calls. The specified value overrides the absence

of the MLIB_NUMBER_OF_THREADS environment variable or any value assigned to it. Refer

to the mlib_setnumthreads(3m) man page for usage information.

To modify the effect of MLIB_NUMBER_OF_THREADS and MLIB_SETNUMTHREADS, you can also

use MP_NUMBER_OF_THREADS at runtime. When MP_NUMBER_OF_THREADS is set to a value

smaller than MLIB_NUMBER_OF_THREADS and MLIB_SETNUMTHREADS, the value of

MP_NUMBER_OF_THREADS takes precedence and limits the amount of parallelism to

MP_NUMBER_OF_THREADS.

These controls set the maximum amount of SMP parallelism that your program can use, and

the VECLIB-specific mechanisms offer finer control within that maximum. Refer to

Performance Benefits below for more information.

Performance Benefits

If MLIB parallelism is enabled, each parallelized MLIB subprogram determines at runtime if

multiple processors are available. If multiple processors are available, MLIB detects whether

the program is already using multiple threads, and uses this information to automatically

choose between a single- or parallel-processor algorithm.

If you are using an HP server with multiple processors, you can realize the performance

benefits of parallel processing in three ways:

1. Call any parallelized MLIB subprogram. Let it use parallelism internally (if it determines

that it is appropriate to do so) based on such factors as problem size, system configuration,

and user environment.