METRIC.ISSUE_EVAL [<ARCHITECTURES>]

The formula used to calculate whether this metric is marked as an issue in the grid and Summary window. When a metric is marked as an issue, it is distinguished as follows:

  • In the Summary window, the metric value is shown with a pink background, and a block of issue text appears under the metric name and value.

  • In the grid, issues are determined on a per-row basis. A cell that is determined to be an issue for a given metric is shown with a pink background, and a tooltip displaying the issue text appears when you mouse over the problem cell.

issue_eval formulas are specified exactly as formulas, however:

  • The function formula() can be used as shorthand to stand for the metric formula attribute for this architecture.

  • The comparison operators (for example, < ) are used to tie the final value to 0 or non-zero. Non-zero final values are marked as issues, zero final values are not. So, for example, in the grid, the following metric will evaluate to a non-zero value if and only if the value shown in the cell ( formula() ) is greater than .3 ( formula() > .3 ), and the row accounts for at least 5% of observed clockticks ( query("PMUHotspot") >.05 ): myMetric.issue_eval["snb"] = ( formula() > .3 ) * ( query("PMUHotspot")> .05 ):

    myMetric.issue_eval["snb"] = ( formula() > .3 ) * ( query("PMUHotspot") > .05 )

    Note

    This attribute must have an architecture specification.

For more complete information about compiler optimizations, see our Optimization Notice.