The following figure shows metric sets for the AI Engine (core) module, the memory module, and the interface module. These metric sets are generally mutually exclusive. You can use a metric from one module type (for example, AI Engine) with one from another type (for example, memory), per your design requirements.
Initially, AMD recommends that you
use the heat_map
metric for the AI Engine module and conflicts metric
for memory modules. Because these settings are specified at runtime, you can run
them as many times as required with your preferences.
Some metrics sets are interconnected; some use group events and others
use individual events. For example, the heat_map
metric set contains a metric that groups all kinds of stall events in a single
metric along with other metrics that group data transfer events (load/store,
streams, cascade, etc,) and vector instructions. Re-run the stalls metric set to get a better view of which stall type(s) are
prevalent. To better understand execution, re-run with the execution metric set.