You can analyze bandwidth data running the following analysis types:
- Bandwidth analysis targeted for Intel® microarchitecture code name Sandy Bridge
General Exploration analysis type targeted for Intel microarchitecture code name Sandy Bridge with the Analyze memory bandwidth option enabled
- Write/Read Bandwidth analysis types targeted for Intel microarchitecture code name Nehalem
When the analysis is complete, the Intel VTune™ Amplifier opens the Bandwidth viewpoint. This viewpoint displays data per memory-access-correlated event-based metrics. Each metric is an event ratio defined by Intel architects and has its own predefined threshold. VTune Amplifier analyzes a ratio value for each aggregated program unit (for example, function). When this value exceeds the threshold and the program unit has more then 5% of CPU time from collection CPU time, it signals a potential performance problem and highlights such a value in pink.
Note
Unlike processors based on the Intel microarchitecture code name Sandy Bridge, the Intel processors code name Beckton or Eagleton only permit either read or write bandwidth to be collected at the same time. So, to see both read and write bandwidth on these processors, perform two collections: one for read and one for write bandwidth.
To interpret the performance data provided during the bandwidth analysis, you may follow the steps below:
Analyze Areas of Significant Memory Bandwidth
To identify bandwidth issues in your application, click the Bottom-up tab and explore the Timeline pane provided at the top of the Bottom-up window. For the Intel microarchitectures code name Sandy Bridge and Nehalem, bandwidth events are not associated with any core, but are instead associated with the uncore (iMC, the integrated memory controller). Uncore events happen on structures shared between all CPUs in a package (for example, 10 CPUs on a single package). This makes it impossible to associate any single uncore event with any code context. So, the VTune Amplifier may only associate bandwidth uncore event counts with the socket, or package, on which the uncore event happened, and time.
For Intel processors code name Beckton or Eagleton based on the Intel microarchitecture code name Nehalem, the Bandwidth viewpoint displays Read Bandwidth and Write Bandwidth data per packages:
For processors based on the Intel microarcitecture code name Nehalem, the Bandwidth viewpoint displays how the Read and Write Bandwidth metrics for your application are changing over time:
Hover over a bar with high bandwidth value to learn how much data was read from or written to DRAM through the on-chip memory controller. Use time-filtering context menu options to filter in a specific range of time during which bandwidth is notable. Then, switch to the core-based events that correlate with bandwidth in the grid below to determine what specific code is inducing all the bandwidth.
Identify Code Sections Inducing Bandwidth
Explore the Bottom-up pane (grid) provided below the Timeline pane. Each row represents a program unit (Function, by default) and percentage of time used by this unit. Program units that take more than 5% of the CPU time are considered as hotspots. By default, the VTune Amplifier sorts the data in the descending order by one of the metrics (CPU Time, for example) and provides the hotspots at the top of the list.
When you select a range of high bandwidth values in the timeline, the VTune Amplifier automatically updates the grid highlighting programs units executed at this time frame. Analyze the performance of these program units per memory-correlated metrics and identify the most critical ones.
Note
To understand the metric, hover over the column header and read the description.
VTune Amplifier analyzes the calculated values, marks those that exceed the threshold set up by Intel architects as performance issues, and highlights them in pink. Hover over a highlighted value to read the issue description:
Analyze Source
When you identified a critical function, double-click it to open the Source/Assembly window and analyze the source code. The Source/Assembly window displays event data. Focus on the events used in the hardware metric identified as performance-critical in the Bottom-up pane (see the calculation formula for the metric). You may sort the columns to locate the required event data leftmost or set the required event column as a Data of Interest. VTune Amplifier remembers your settings and restores them each time you open your result.
Note
For information on processor event, see the Reference for Processor Events book in this online help.
For information on the performance tuning for HPC-computers using the event-based sampling collection, see http://software.intel.com/en-US/articles/processor-specific-performance-analysis-papers/.
See Also
Supplemental documentation specific to a particular Intel Studio may be available at <install-dir>\<studio>\documentation\ .