Mentor v8.6_4 manual ∙ patternfilename, ∙ -multiplemacros macrofilename

Command Dictionary Macrotest

∙instance_name

A non-repeatable string that specifies the name of the instance to use in the macrotest.

∙pattern_file_name

The set of patterns to be applied to the macro.

∙-multiple_macros macro_filename

An optional switch that allows multiple macros to be tested in parallel, by allowing a file containing multiple macrotest commands, macro_filename, one command for each of the macros to be tested in parallel. Each macro can be tested individually, in a separate run. Once each macro is successfully tested, its corresponding command can simply be removed from the dofile (or command line) and placed in this macro_filename, which can only contain macrotest commands.

If the -multiple_macros switch is used, all macros to be simultaneously tested must be included in the file. It is not possible to use this switch and also specify a macro instance name or pattern file on the same command line. It is possible to test one subset of all macros using one -multiple_macros run, with one macro_filename containing those macros and their pattern file specifications, and then test another subset by using another -multiple_macros run. If options appear with the -multiple_macros switch, then those options are used for all macros in the file. The following example:

macrotest -multiple_macros macrofile -random -max_path_attempts 20

causes all of the macros to be tested in parallel (those with a macrotest command in file “ macrofile” ) to use the -random option with at most 20 observation attempts per macro output per pattern.

∙-FIll_patterns

An optional literal that causes unspecified values in the scan patterns created by macrotest to be randomly filled. This is the default.

Typically, only a small number of the scan chain bits need to be specified to deliver tests to the macro. The remaining bits of the scan chain are unspecified

(X). This can cause false simulation messages (such as possible bus contention) because ATPG has intelligence missing from simulators. By filling known values, the simulator is forced to understand that no problem exists.

FastScan and FlexTest Reference Manual, V8.6_4 2-243