jkoshy 1d3209ab83 MFP4:
- Implement sampling modes and logging support in hwpmc(4).

- Separate MI and MD parts of hwpmc(4) and allow sharing of
  PMC implementations across different architectures.
  Add support for P4 (EMT64) style PMCs to the amd64 code.

- New pmcstat(8) options: -E (exit time counts) -W (counts
  every context switch), -R (print log file).

- pmc(3) API changes, improve our ability to keep ABI compatibility
  in the future.  Add more 'alias' names for commonly used events.

- bug fixes & documentation.
2005-06-09 19:45:09 +00:00

235 lines
7.0 KiB
Groff

.\" Copyright (c) 2003 Joseph Koshy. All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" This software is provided by Joseph Koshy ``as is'' and
.\" any express or implied warranties, including, but not limited to, the
.\" implied warranties of merchantability and fitness for a particular purpose
.\" are disclaimed. in no event shall Joseph Koshy be liable
.\" for any direct, indirect, incidental, special, exemplary, or consequential
.\" damages (including, but not limited to, procurement of substitute goods
.\" or services; loss of use, data, or profits; or business interruption)
.\" however caused and on any theory of liability, whether in contract, strict
.\" liability, or tort (including negligence or otherwise) arising in any way
.\" out of the use of this software, even if advised of the possibility of
.\" such damage.
.\"
.\" $FreeBSD$
.\"
.Dd Dec 15, 2003
.Os
.Dt PMCSTAT 8
.Sh NAME
.Nm pmcstat
.Nd performance measurement with performance monitoring hardware
.Sh SYNOPSIS
.Nm
.Op Fl D Ar pathname
.Op Fl C
.Op Fl E
.Op Fl O Ar logfilename
.Op Fl P Ar event-spec
.Op Fl R Ar logfilename
.Op Fl S Ar event-spec
.Op Fl W
.Op Fl c Ar cpu
.Op Fl d
.Op Fl g
.Op Fl m
.Op Fl n Ar count
.Op Fl o Ar outputfile
.Op Fl p Ar event-spec
.Op Fl s Ar event-spec
.Op Fl t Ar pid
.Op Fl w Ar interval
.Op command Op args
.Sh DESCRIPTION
The
.Nm
utility measures system performance using the facilities provided by
.Xr hwpmc 4 .
.Pp
The
.Nm
utility can measure both hardware events seen by the system as a
whole, and those seen when a specified process is executing on the
system's CPUs.
If a specific process is being targeted (for example,
if the
.Fl t Ar pid
option is specified, or if a command line is specified using
.Ar command ) ,
then measurement occurs till the target process exits or
the
.Nm
utility is interrupted by the user.
If a specific process is not targeted for measurement, then
.Nm
will perform system-wide measurements till interrupted by the
user.
.Pp
A given invocation of
.Nm
can mix allocations of system-mode and process-mode PMCs, of both
counting and sampling flavors.
The values of all counting PMCs are printed in human readable form
at regular intervals by
.Nm .
The output of sampling PMCs may be configured to go to a log file for
subsequent offline analysis, or, at the expense of greater
overhead, may be configured to be processed on the fly.
.Pp
Hardware events to measure are specified to
.Nm
using event specifier strings
.Ar event-spec .
The syntax of these event specifiers is machine dependent and is
documented in
.Xr pmc 3 .
.Pp
A process-mode PMC may be configured to be inheritable by the target
process' current and future children.
.Sh OPTIONS
The following options are available:
.Bl -tag -width indent
.It Fl C
Toggle between showing cumulative or incremental counts for
subsequent counting mode PMCs specified on the command line.
The default is to show incremental counts.
.It Fl D Ar pathname
Create files with per-program samples in the directory named
by
.Ar pathname .
The default is to create these files in the current directory.
.It Fl E
Toggle showing per-process counts at the time a tracked process
exits for subsequent process-mode PMCs specified on the command line.
This option is useful for mapping the performance characteristics of a
complex pipeline of processes when used in conjunction with the
.Fl d
option.
The default is to not to enable per-process tracking.
.It Fl O Ar logfilename
Send the output of sampling mode PMCs to
.Ar logfilename .
If this option is not specified and one of the logging options
is requested, then
.Nm
will print a human-readable version of the log to the configured
output file.
.It Fl P Ar event-spec
Allocate a process mode sampling PMC measuring hardware events
specified in
.Ar event-spec .
.It Fl R Ar logfilename
Perform offline analysis using sampling data in file
.Ar logfilename .
.It Fl S Ar event-spec
Allocate a system mode sampling PMC measuring hardware events
specified in
.Ar event-spec .
.It Fl W
Toggle logging the incremental counts seen by the threads of a
tracked process each time they are scheduled on a CPU.
This is an experimental feature intended to help analyse the
dynamic behaviour of processes in the system.
It may incur substantial overhead if enabled.
The default is for this feature to be disabled.
.It Fl c Ar cpu
Set the cpu for subsequent system mode PMCs specified on the
command line to
.Ar cpu .
The default is to allocate system mode PMCs on CPU zero.
.It Fl d
Toggle between process mode PMCs measuring events for the target
process' current and future children or only measuring events for
the target process.
The default is to measure events for the target process alone.
.It Fl g
Produce execution profiles in a format compatible with
.Xr gprof 1 .
.It Fl m
When producing
.Xr gprof 1
compatible execution profiles, merge profiles across multiple
invocations of the same executable.
.It Fl n Ar rate
Set the default sampling rate for subsequent sampling mode
PMCs specified on the command line.
The default is to configure PMCs to sample the CPU's instruction
pointer every 65536 events.
.It Fl o Ar outputfile
Send counter readings and printed representations of logged data
to file
.Ar outputfile .
The default is to send output to
.Pa stderr .
.It Fl p Ar event-spec
Allocate a process mode counting PMC measuring hardware events
specified in
.Ar event-spec .
.It Fl s Ar event-spec
Allocate a system mode counting PMC measuring hardware events
specified in
.Ar event-spec .
.It Fl t Ar pid
Attach all process mode PMCs allocated to the process with PID
.Ar pid .
The option is not allowed in conjunction with specifying a
command using
.Ar command .
.It Fl w Ar secs
Print the values of all counting mode PMCs every
.Ar secs
seconds.
The argument
.Ar secs
may be a fractional value.
The default interval is 5 seconds.
.El
.Pp
If
.Ar command
is specified, it is executed using
.Xr execvp 3 .
.Sh EXAMPLES
To perform system-wide statistical sampling on an AMD Athlon CPU with
samples taken every 32768 instruction retirals and data being sampled
to file
.Dq sample.stat ,
use:
.Dl pmccstat -O sample.stat -n 32768 -S k7-retired-instructions
.Pp
To execute
.Dq mozilla
and measure the number of data cache misses suffered
by it and its children every 12 seconds on an AMD Athlon, use:
.Dl pmcstat -d -w 12 -p k7-dc-misses mozilla
.Sh DIAGNOSTICS
.Ex -std pmcstat
.Sh HISTORY
The
.Nm
utility is proposed to be integrated into
.Fx
sometime after
.Fx 5.2 .
.Nm
.Bt
.Sh AUTHORS
.An Joseph Koshy Aq jkoshy@FreeBSD.org
.Sh SEE ALSO
.Xr gprof 1 ,
.Xr execvp 3 ,
.Xr pmc 3 ,
.Xr hwpmc 4 ,
.Xr pmccontrol 8 ,
.Xr sysctl 8