Semantic Grouping for Program Performance Data Analysis

A computer-readable non-transitory storage medium configured with data and with instructions that when executed by at least one processor causes the processor(s) to perform a process for organizing program performance data, the process comprising the steps of:

• obtaining performance data which includes multiple samples, each of the samples having a stack of names which represent nodes located in a directed acyclic graph (DAG), each of the nodes having an associated cost;
• receiving a grouping expression, written in a transformation syntax language, to group nodes and consolidate costs by following at least one of the following procedures:
• a clustering-by-name procedure,
• an entry-group-clustering procedure,
• a folding-by-name procedure,
• a folding-by-cost procedure;
• grouping nodes into at least one group as commanded in the grouping expression;
• consolidating costs by associating with each group a total cost which is the sum of the cost(s) associated with individual node(s) of the group;
• and
• producing a cost accounting showing group name(s) and associated group costs after grouping nodes and consolidating costs.

A computer system comprising:

• a logical processor;
• a memory in operable communication with the logical processor;
• performance data residing in the memory and having multiple samples, each of the samples having a stack of names which represent nodes located in a directed acyclic graph (DAG), each of the nodes having an associated cost;
• and
• a modified stack residing in the memory, the modified stack having at least a portion of at least one name and/or at least one name-associated characteristic in common with a stack of the performance data, the modified stack being different from the performance data in that at least one of the following is present in the modified stack and absent from the performance data:
• a clustering-by-name group,
• an entry-group-clustering group,
• a folding-by-name group,
• a folding-by-cost group.