Parallel program profiler

A non-transitory computer-readable medium for storing instructions, the instructions comprising:

• one or more instructions which, when executed by at least one processor, cause the at least one processor to:
• receive information generated by execution of a parallel technical computing environment (TCE)-based program,
• two or more units of execution (UEs) executing portions of the parallel TCE-based program;
• generate a plurality of performance indicators for the parallel TCE-based program on one of a line basis or a function basis and based on the received information;
• create an array of information for each executed line of the parallel TCE-based program executed by each of the two or more UEs;
• and
• provide, for display, a single interface that includes information regarding the plurality of performance indicators for each UE and a comparative view, based on the created array of information for each executed line, of a plurality of executed lines of the parallel TCE-based program,
• each executed line, of the plurality of executed lines of the parallel program, being associated with at least one of the plurality of performance indicators,
• each executed line, of the plurality of executed lines of the parallel program, including information associated with each of the two or more UEs, and
• each executed line being provided, for display, in a corresponding row of the single interface,
• a row including first information including two or more of the plurality of performance indicators and second information including two or more of the plurality of performance indicators,
• the first information being further associated with a first UE, of the two or more UEs, and
• the second information being further associated with a second UE, of the two or more UEs.

A system comprising:

• one or more devices, comprised at least partially of hardware, to:
• receive information generated by execution of a parallel program by a technical computing environment (TCE),
• two or more units of execution (UEs) executing portions of the parallel program,
• generate a plurality of performance indicators for the parallel program on one of a line basis or a function basis and based on the received information,
• create an array of information for each executed line of the parallel program executed by each of the two or more UEs;
• and
• provide a user interface that displays an interface that simultaneously includes information regarding the plurality of performance indicators for each UE and a comparative view, based on the created array of information for each executed line, of a plurality of executed lines of the parallel program,
• each executed line, of the plurality of executed lines of the parallel program, being associated with at least one of the plurality of performance indicators,
• each executed line, of the plurality of executed lines of the parallel program, including information associated with each of the two or more UEs, and
• each executed line being provided, for display, in a corresponding row of the interface,
• a row including first information including two or more of the plurality of performance indicators and second information including two or more of the plurality of performance indicators,
• the first information being further associated with a first UE, of the two or more UEs, and
• the second information being further associated with a second UE, of the two or more UEs.

A device comprising:

• a memory to store instructions;
• and
• a processor to execute the instructions to:
• receive information generated by execution of a parallel program by a technical computing environment (TCE),
• two or more units of execution (UEs) executing portions of the parallel program;
• aggregate the received information;
• generate a plurality of performance indicators for the parallel program on one of a line basis or a function basis and based on the aggregated information;
• create an array of information for each executed line of the parallel program executed by each of the two or more UEs;
• and
• provide, for display, a single interface that includes information regarding the plurality of performance indicators for each UE and a comparative view, based on the created array of information for each executed line, of a plurality of executed lines of the parallel program to a user,
• each executed line, of the plurality of executed lines of the parallel program, being associated with at least one of the plurality of performance indicators,
• each executed line, of the plurality of executed lines of the parallel program, including information associated with each of the two or more UEs, and
• each executed line being provided, for display, in a corresponding row of the single interface,
• a row including first information including two or more of the plurality of performance indicators and second information including two or more of the plurality of performance indicators,
• the first information being further associated with a first UE, of the two or more UEs, and
• the second information being further associated with a second UE, of the two or more UEs.