Techniques for Optimizing CPU Performance

Reducing CPU Time by Using More Memory or Reducing I/O

Executing a single stream of code takes approximately the same amount of CPU time each time that code is executed. Optimizing CPU performance in these instances is usually a tradeoff . For example, you might be able to reduce CPU time by using more memory, because more information can be read and stored in one operation, but less memory is available to other processes.

Also, because the CPU performs all the processing that is needed to perform an I/O operation, an option or technique that reduces the number of I/O operations can also have a positive effect on CPU usage.

Storing a Compiled Program for Computation- Intensive DATA Steps

Another technique that can improve CPU performance is to store a DATA step that is executed repeatedly as a compiled program rather than as SAS statements. This is especially true for large DATA step jobs that are not I/O-intensive. For more information on storing compiled DATA steps, see Chapter 30, 'Stored Compiled DATA Step Programs,' on page 547.

Reducing Search Time for SAS Executable Files

The PATH= system option specifies the list of directories (or libraries, in some operating environments) that contain SAS executable files. Your default configuration file specifies a certain order for these directories. You can rearrange the directory specifications in the PATH= option so that the most commonly accessed directories are listed first. Place the least commonly accessed directories last.

Operating Environment Information: The PATH= system option is not available in some operating environments. See the documentation for your operating environment for more information.

Specifying Variable Lengths

When SAS processes the program data vector, it typically moves the data in one large operation rather than by individual variables . When data is properly aligned (in 8-byte boundaries), data movement can occur in as little as 2 clock cycles (a single load followed by a single store). SAS moves unaligned data by more complex means, at worst, a single byte at a time. This would be at least eight times slower for an 8-byte variable.

Many high performance RISC (Reduced Instruction Set Computer) processors pay a very large performance penalty for movement of unaligned data. When possible, leave numeric data at full width (eight bytes). Note that SAS must widen short numeric data for any arithmetic operation. On the other hand, short numeric data can save both memory and I/O. You must determine which method is most advantageous for your operating environment and situation.

Note: Alignment can be especially important when you process a data set by selecting only specific variables or when you use WHERE processing.

Using Parallel Processing

SAS System 9 supports a new wave of SAS functionality related to parallel processing. Parallel processing means that processing is handled by multiple CPUs simultaneously . This technology takes advantage of SMP machines and provides performance gains for two types of SAS processes: threaded I/O and threaded application processing.

For information, see Chapter 14, 'Support for Parallel Processing,' on page 223.