Scroll to navigation

LIBPFM(3) Linux Programmer's Manual LIBPFM(3)

NAME

libpfm_core - support for Intel Atom processor family

SYNOPSIS

#include <perfmon/pfmlib.h>
#include <perfmon/pfmlib_intel_atom.h>

DESCRIPTION

The libpfm library provides full support for the Intel Atom processor. This processor implements Intel architectural perfmon v3 with Precise Event-Based Sampling (PEBS) support. It also implements all architected events to which it adds lots of Atom specific events.

The libpfm interface is defined in pfmlib_intel_atom.h. It consists of a set of functions and structures which describe and allow access to the Intel Atom processor specific PMU features.

When Intel Atom processor specific features are needed to support a measurement, their descriptions must be passed as model-specific input arguments to the pfm_dispatch_events() function. The Intel Atom processors specific input arguments are described in the pfmlib_intel_atom_input_param_t structure. No output parameters are currently defined. The input parameters are defined as follows:

typedef struct {
	unsigned int	cnt_mask;
	unsigned int	flags;
} pfmlib_intel_atom_counter_t;
typedef struct {

pfmlib_intel_atom_counter_t pfp_intel_atom_counters[PMU_INTEL_ATOM_NUM_COUNTERS];
unsigned int pfp_intel_atom_pebs_used;
uint64_t reserved[4]; } pfmlib_core_input_param_t;

The Intel Atom processor provides several additional per-event features for counters: thresholding, inversion, edge detection, monitoring both threads. They can be set using the pfp_intel_atom_counters data structure for each event. The flags field can be initialized with any combinations of the following values:

Inverse the results of the cnt_mask comparison when set
Enable edge detection of events.
Enable measuring the event in any of the two threads. By default only the current thread is measured.

The cnt_mask field is used to set the event threshold. The value of the counter is incremented each time the number of occurrences per cycle of the event is greater or equal to the value of the field. Thus the event is modified to actually measure the number of qualifying cycles. When zero all occurrences are counted (this is the default).

Support for Precise-Event Based Sampling (PEBS)

The library can be used to setup the PMC registers when using PEBS. In this case, the pfp_intel_atom_pebs_used field must be set to 1. When using PEBS, it is not possible to use more than one event.

ERRORS

Refer to the description of the pfm_dispatch_events() function for errors.

SEE ALSO

pfm_dispatch_events(3) and set of examples shipped with the library

AUTHOR

Stephane Eranian <eranian@gmail.com>

November, 2006