linux-toradex.git/tools/perf/bench/numa.c, branch v4.1

perf bench numa: Fix immediate meeting of convergence condition

2015-04-27T16:57:50+00:00

This patch fixes the race in the beginning of benchmark run when some
threads hasn't got assigned curr_cpu yet so they don't occur in
nodes-of-process stats and benchmark concludes that all remaining
threads are converged already.

The race can be reproduced with small amount of threads and some bigger
amount of shared process memory, e.g. one process, two threads and 5GB
of process memory.

Signed-off-by: Petr Holasek 
Reviewed-by: Ingo Molnar 
Cc: Jiri Olsa 
Link: http://lkml.kernel.org/r/1429198699-25039-4-git-send-email-pholasek@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo

perf bench numa: Fixes of --quiet argument

2015-04-27T16:57:49+00:00

Corrected description and fixed function of --quiet argument.

Signed-off-by: Petr Holasek 
Reviewed-by: Ingo Molnar 
Cc: Jiri Olsa 
Link: http://lkml.kernel.org/r/1429198699-25039-2-git-send-email-pholasek@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo

Merge branch 'perf-core-for-mingo' into perf/urgent

2014-04-14T14:45:39+00:00

Conflicts:
	tools/perf/bench/numa.c

Pull perf fixes from Jiri Olsa.

Signed-off-by: Ingo Molnar

perf bench: Set more defaults in the 'numa' suite

2014-04-14T10:55:58+00:00

Currently,

  $ perf bench numa mem

errors out with usage information. To make this more user-friendly, let
us provide a minimum set of default values required for a test
run. As an added bonus,

  $ perf bench all

now goes all the way to completion.

Signed-off-by: Ramkumar Ramachandra 
Cc: Arnaldo Carvalho de Melo 
Cc: David Ahern 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Link: http://lkml.kernel.org/r/1395964219-22173-2-git-send-email-artagnon@gmail.com
Signed-off-by: Jiri Olsa

perf bench numa: Make no args mean 'run all tests'

2014-03-14T13:04:10+00:00

If we call just:

  perf bench numa mem

it will present the same output as:

  perf bench numa mem -h

i.e. ask for instructions about what to run.

While that is kinda ok, using 'run all tests' as the default, i.e.
making 'no parms' be equivalent to:

  perf bench numa mem -a

Will allow:

  perf bench numa all

to actually do what is asked: i.e. run all the 'bench' tests, instead of
responding to that by asking what to do.

That, in turn, allows:

  perf bench all

to actually complete, for the same reasons.

And after that, the tests that come after that, and that at some point
hit a NULL deref, will run, allowing me to reproduce a recently reported
problem.

That when you have the needed numa libraries, which wasn't the case for
the reporter, making me a bit confused after trying to reproduce his
report.

So make no parms mean -a.

Cc: Adrian Hunter 
Cc: David Ahern 
Cc: Frederic Weisbecker 
Cc: Jiri Olsa 
Cc: Mike Galbraith 
Cc: Namhyung Kim 
Cc: Patrick Palka 
Cc: Paul Mackerras 
Cc: Peter Zijlstra 
Cc: Stephane Eranian 
Link: http://lkml.kernel.org/n/tip-x7h0ghx4pef4n0brywg21krk@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Fix bench/numa.c for 32-bit build

2013-10-21T14:19:42+00:00

bench/numa.c: In function 'worker_thread':
bench/numa.c:1123:20: error: comparison between signed and unsigned integer expressions [-Werror=sign-compare]
bench/numa.c:1171:6: error: format '%lx' expects argument of type 'long unsigned int', but argument 5 has type 'u64' [-Werror=format]
cc1: all warnings being treated as errors

Signed-off-by: Adrian Hunter 
Cc: David Ahern 
Cc: Frederic Weisbecker 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Mike Galbraith 
Cc: Namhyung Kim 
Cc: Paul Mackerras 
Cc: Peter Zijlstra 
Cc: Stephane Eranian 
Link: http://lkml.kernel.org/r/1382099356-4918-13-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo

perf bench: Fix failing assertions in numa bench

2013-10-11T15:17:38+00:00

Patch adds more subtle handling of -C and -N parameters in
parse_{cpu,node}_setup_list() functions when there isn't enough NUMA
nodes or CPUs present.  Instead of assertion and terminating benchmark,
partial test is skipped with error message and perf will continue to the
next one.

Fixed problem can be easily reproduced on machine with only one NUMA
node:

 # Running numa/mem benchmark...

  # Running main, "perf bench numa mem -a"

...

 # Running RAM-bw-remote, "perf bench numa mem -p 1 -t 1 -P 1024 -C 0 -M 1 -s
perf: bench/numa.c:622: parse_setup_node_list: Assertion `!(bind_node_0 < 0 ||
		bind_node_0 >= g->p.nr_nodes)' failed.
Aborted

Signed-off-by: Petr Holasek 
Acked-by: Ingo Molnar 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Petr Benas 
Link: http://lkml.kernel.org/r/1380821325-4017-1-git-send-email-pholasek@redhat.com
Signed-off-by: Petr Benas 
Signed-off-by: Arnaldo Carvalho de Melo

perf: Add 'perf bench numa mem' NUMA performance measurement suite

2013-01-30T13:35:36+00:00

Add a suite of NUMA performance benchmarks.

The goal was simulate the behavior and access patterns of real NUMA
workloads, via a wide range of parameters, so this tool goes well
beyond simple bzero() measurements that most NUMA micro-benchmarks use:

 - It processes the data and creates a chain of data dependencies,
   like a real workload would. Neither the compiler, nor the
   kernel (via KSM and other optimizations) nor the CPU can
   eliminate parts of the workload.

 - It randomizes the initial state and also randomizes the target
   addresses of the processing - it's not a simple forward scan
   of addresses.

 - It provides flexible options to set process, thread and memory
   relationship information: -G sets "global" memory shared between
   all test processes, -P sets "process" memory shared by all
   threads of a process and -T sets "thread" private memory.

 - There's a NUMA convergence monitoring and convergence latency
   measurement option via -c and -m.

 - Micro-sleeps and synchronization can be injected to provoke lock
   contention and scheduling, via the -u and -S options. This simulates
   IO and contention.

 - The -x option instructs the workload to 'perturb' itself artificially
   every N seconds, by moving to the first and last CPU of the system
   periodically. This way the stability of convergence equilibrium and
   the number of steps taken for the scheduler to reach equilibrium again
   can be measured.

 - The amount of work can be specified via the -l loop count, and/or
   via a -s seconds-timeout value.

 - CPU and node memory binding options, to test hard binding scenarios.
   THP can be turned on and off via madvise() calls.

 - Live reporting of convergence progress in an 'at glance' output format.
   Printing of convergence and deconvergence events.

The 'perf bench numa mem -a' option will start an array of about 30
individual tests that will each output such measurements:

 # Running  5x5-bw-thread, "perf bench numa mem -p 5 -t 5 -P 512 -s 20 -zZ0q --thp  1"
  5x5-bw-thread,                         20.276, secs,           runtime-max/thread
  5x5-bw-thread,                         20.004, secs,           runtime-min/thread
  5x5-bw-thread,                         20.155, secs,           runtime-avg/thread
  5x5-bw-thread,                          0.671, %,              spread-runtime/thread
  5x5-bw-thread,                         21.153, GB,             data/thread
  5x5-bw-thread,                        528.818, GB,             data-total
  5x5-bw-thread,                          0.959, nsecs,          runtime/byte/thread
  5x5-bw-thread,                          1.043, GB/sec,         thread-speed
  5x5-bw-thread,                         26.081, GB/sec,         total-speed

See the help text and the code for more details.

Cc: Peter Zijlstra 
Cc: Arnaldo Carvalho de Melo 
Cc: Frederic Weisbecker 
Cc: Mike Galbraith 
Cc: Steven Rostedt 
Cc: Linus Torvalds 
Cc: Andrew Morton 
Cc: Peter Zijlstra 
Cc: Andrea Arcangeli 
Cc: Rik van Riel 
Cc: Mel Gorman 
Cc: Hugh Dickins 
Signed-off-by: Ingo Molnar