Application Summary: pcluster_accel


Static Analysis


The following table shows static analysis metrics ( Cyclomatic Complexity, programming languages, types of parallelism, lines of code, etc.)

Click on a column name to sort. Hover over the Cyclomatic Complexity value for details.
Caveat: This metric may be inaccurate if the program uses languages other than C/C++.


App Version Web Domain Origin Prog. Lang Parallelism LOC #Files #Functions Cyclomatic Complexity Details
pcluster_accel 1.0 Web Climate Data Analytics ORNL C MPI 1937 2 43
301
Details

Instruction Mix Table


The following table shows the distribution of instructions (percentages) for an application across various categories.
It is dynamically generated from live performance data in PADS.

This information is obtained using a tool based on Intel's PIN, a dynamic binary instrumentation tool. We discard NOP and Prefetch instructions while calculating total instructions.

The table heatmap is generated by categorizing a cell's value into six categories(represented by different colors). So, every cell is colored based on the magnitude of the value it contains.
Higher magnitudes are denoted by darker colors.

Click anywhere on a row to select for comparison.
Click on a column name to sort.
Click on the application name to see a summary chart.
App Exp Platform BrOpsFlOpsFp SIMDInt SIMDIntOpsMem SIMDMemOpsMiscMoves Details
pcluster_accel master-kernel-tnnc_gsmnp_vertical_profiles-1.7GB-16p Optimus 19.560.000.000.0233.600.0143.220.013.58 Details
pcluster_accel slave-kernel-tnnc_gsmnp_vertical_profiles-1.7GB-16p Optimus 14.2828.200.230.1226.670.5728.240.011.69 Details
pcluster_accel master-overall-tnnc_gsmnp_vertical_profiles-1.7GB-16p Optimus 19.510.000.000.0233.660.0143.140.013.64 Details
pcluster_accel slave-overall-tnnc_gsmnp_vertical_profiles-1.7GB-16p Optimus 14.0529.190.240.1226.530.6027.650.011.63 Details


Instruction Mix Comparison Chart

Summary (% total of instructions by type)

Application: Experiment
    pcluster_accel : master-kernel-tnnc_gsmnp_vertical_profiles-1.7GB-16p
    pcluster_accel : slave-kernel-tnnc_gsmnp_vertical_profiles-1.7GB-16p
    pcluster_accel : master-overall-tnnc_gsmnp_vertical_profiles-1.7GB-16p
    pcluster_accel : slave-overall-tnnc_gsmnp_vertical_profiles-1.7GB-16p



Communication Behavior

The following figure provides a preview of the communication pattern of the application.
It is dynamically generated from live performance data in PADS.

An application's message volume and counts are obtained using mpiP, a lightweight profiling library for MPI applications. It is used to collect basic communication characteristics (point-to-point and collective calls) as well as the communication topology.



pcluster_accel - tnnc_gsmnp_vertical_profiles-1.7GB-64p

pcluster_accel - tnnc_gsmnp_vertical_profiles-1.7GB
.


Memory Bandwidth Data


The following table shows the memory bandwidth information for the following applications.
It is dynamically generated from live performance data in PADS.

This information is obtained using the memory bandwidth tool that is based on PAPI hardware counters. The statistics represent the mean value for each metric across the different tasks.



App Num Tasks Exp BW_Read_CPU(B/Cycle)BW_Read_Wall(MB/s)BW_Write_CPU(B/Cycle)BW_Write_Wall(MB/s)CpuCyclesMemReadsMemWritesWallClock(us) Details
pcluster_accel 16 kernel-tnnc_gsmnp_vertical_profiles-1.7GB-16p 0.08121.550.0458.8655064564863.4765108531.9431717634.3534154947.06 Details