Kernel Execution Input

Developer Guide for Intel® SDK for OpenCL™ Applications 2017

Download PDF

ID 773042

Date 10/22/2018

Version

Public

A newer version of this document is available. Customers should click here to go to the newest version.

Visible to Intel only — GUID: GUID-D655A00E-2C6D-43BF-862F-61454C95B342

To assign analysis inputs for kernel execution, do the following:

Select the desirable kernel from the session's kernels list in the Code Builder Session Explorer or from the Select Kernel combo box in the Code Builder Analyze toolbar.
Open the Code Builder Analysis Input window from:
In Visual Studio*: Tools > Intel Code Builder for OpenCL API > OpenCL Kernel Development > Windows > Analysis Input
or by clicking the Open Analysis Input button in the Code Builder Analysis toolbar.
Assign a variable for each kernel argument in the Kernel Arguments table by clicking the Click here to assign link under the Assigned Variable column. You can assign one-dimensional variables (such as integer, float, char, half, and so on) on-the-fly by typing single values into the table. See Variable Management for details.
You can assign a reference for each output variable (buffers or images) that you want to verify. If you specify a validation reference, after the kernel execution each one of the output variables will be compared with its assigned reference in order to validate the correctness of the kernel executions.
Set the number of iterations, global size and local sizes per workload dimension in the Workgroup size definitions group box.

You can use the local size(s) text boxes for several different test configurations:

Set single size value for a single test.
Add several comma-separated sizes for multiple tests.
Set 0 to use the default framework-assigned local size.
Check the Auto box to enable the Kernel Builder to iterate on all sizes that are smaller than global size and device maximum local size.

Also consider the following:

Each option is available for each dimension.
To analyze the kernel in its designed conditions, set a single value.
To find the local size that provides higher performance results, click Auto or set a list of comma-separated values.
To improve the analysis accuracy, run each global and local work size combination several times by increasing the Number of iterations value. Several iterations minimize the impact of other system processes or tasks on the kernel execution time.
Use the Device Information dialog to compare device properties and choose the appropriate device for the kernel.

Follow these steps to create a new configuration for an existing built session:

In the Code Builder Analysis Input pane, you need to assign variables for each kernel argument that is listed. Click on the link in the Assigned Variable column to assign the appropriate variable.
Enter a valid name to enable the New Configuration button . Click the button to duplicate the current configuration. You can edit the new configuration to create multiple analysis configurations.