public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Yao Qi <yao@codesourcery.com>
To: <gdb-patches@sourceware.org>
Subject: [RFC 0/3] GDB Performance testing
Date: Wed, 28 Aug 2013 04:17:00 -0000	[thread overview]
Message-ID: <1377663394-4975-1-git-send-email-yao@codesourcery.com> (raw)
In-Reply-To: <520B7F70.6070207@codesourcery.com>

This patch series implement GDB performance testing infrastructure
according to the design I posted here
<https://sourceware.org/ml/gdb-patches/2013-08/msg00380.html>

Here are some highlights:

  - Performance testing can be run via 'make check-perf'
  - GDB and GDBServer is started by dejagnu, so the usage of
'make check-perf' is same as the usage of existing 'make check'.
  - Performance test result is saved in testsuite/perftest.log, which
is appended in multiple runs.
  - Workload of each test can be customized by passing parameters to
'make check-perf'.

The basic usages and the outputs are as follows:

$ make check-perf
$ cat ./testsuite/perftest.log
solib 8 0.01
solib 16 0.03
solib 32 0.07
solib 64 0.19
solib 128 0.54

$ make check-perf RUNTESTFLAGS="--target_board=native-gdbserver solib.exp"
$ cat ./testsuite/perftest.log
solib 8 0.03
solib 16 0.05
solib 32 0.11
solib 64 0.26
solib 128 0.78

Specify the number of solibs in the test.
$ make check-perf RUNTESTFLAGS="--target_board=native-gdbserver solib.exp SOLIB_NUMBER=1024"
$ cat ./testsuite/perftest.log
solib 64 0.25
solib 128 0.7
solib 256 2.38
solib 512 9.67
solib 1024 53.0

GDB python doesn't know the perftest package located in
testsuite/gdb.perf/lib, so in every test, we need the following two
statements to add the path to sys.path.

  gdb_test_no_output "python import os, sys"
  gdb_test_no_output "python sys.path.insert\(0, os.path.abspath\(\"${srcdir}/${subdir}/lib\"\)\)"

I'll add other test cases when the basic form of test case is
determined.

*** BLURB HERE ***

Yao Qi (3):
  New make target 'check-perf' and new dir gdb.perf
  Perf test framework
  Test on solib load and unload

 gdb/Makefile.in                                   |    8 ++
 gdb/testsuite/Makefile.in                         |   15 +++-
 gdb/testsuite/configure                           |    3 +-
 gdb/testsuite/configure.ac                        |    2 +-
 gdb/testsuite/gdb.perf/Makefile.in                |   15 ++++
 gdb/testsuite/gdb.perf/lib/perftest/config.py     |   40 ++++++++++
 gdb/testsuite/gdb.perf/lib/perftest/perftest.py   |   49 ++++++++++++
 gdb/testsuite/gdb.perf/lib/perftest/reporter.py   |   38 +++++++++
 gdb/testsuite/gdb.perf/lib/perftest/testresult.py |   42 ++++++++++
 gdb/testsuite/gdb.perf/solib.c                    |   79 +++++++++++++++++++
 gdb/testsuite/gdb.perf/solib.exp                  |   86 +++++++++++++++++++++
 gdb/testsuite/gdb.perf/solib.py                   |   48 ++++++++++++
 12 files changed, 420 insertions(+), 5 deletions(-)
 create mode 100644 gdb/testsuite/gdb.perf/Makefile.in
 create mode 100644 gdb/testsuite/gdb.perf/lib/perftest/__init__.py
 create mode 100644 gdb/testsuite/gdb.perf/lib/perftest/config.py
 create mode 100644 gdb/testsuite/gdb.perf/lib/perftest/perftest.py
 create mode 100644 gdb/testsuite/gdb.perf/lib/perftest/reporter.py
 create mode 100644 gdb/testsuite/gdb.perf/lib/perftest/testresult.py
 create mode 100644 gdb/testsuite/gdb.perf/solib.c
 create mode 100644 gdb/testsuite/gdb.perf/solib.exp
 create mode 100644 gdb/testsuite/gdb.perf/solib.py

-- 
1.7.7.6

  parent reply	other threads:[~2013-08-28  4:17 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-08-14 13:01 [RFC] GDB performance testing infrastructure Yao Qi
2013-08-21 20:39 ` Tom Tromey
2013-08-27  6:21   ` Yao Qi
2013-08-27 13:49 ` Agovic, Sanimir
2013-08-28  3:04   ` Yao Qi
2013-09-19  0:36     ` Doug Evans
2013-08-28  4:17 ` Yao Qi [this message]
2013-08-28  4:17   ` [RFC 1/3] New make target 'check-perf' and new dir gdb.perf Yao Qi
2013-08-28  9:40     ` Agovic, Sanimir
2013-09-19 17:47     ` Doug Evans
2013-09-20 19:00       ` Tom Tromey
2013-09-20 18:59     ` Tom Tromey
2013-08-28  4:17   ` [RFC 2/3] Perf test framework Yao Qi
2013-08-28  9:57     ` Agovic, Sanimir
2013-09-03  1:45       ` Yao Qi
2013-09-03  6:38         ` Agovic, Sanimir
2013-09-19 19:09     ` Doug Evans
2013-09-20  8:04       ` Yao Qi
2013-09-20 16:51         ` Doug Evans
2013-09-22  2:54           ` Yao Qi
2013-09-22 23:14             ` Doug Evans
2013-09-20 17:12         ` Doug Evans
2013-08-28  4:17   ` [RFC 3/3] Test on solib load and unload Yao Qi
2013-08-28  4:27     ` Yao Qi
2013-08-28 11:31       ` Agovic, Sanimir
2013-09-03  1:59         ` Yao Qi
2013-09-03  6:33           ` Agovic, Sanimir
2013-09-02 15:24       ` Blanc, Nicolas
2013-09-03  2:04         ` Yao Qi
2013-09-03  7:50           ` Blanc, Nicolas
2013-09-19 22:45       ` Doug Evans
2013-09-20 19:19         ` Tom Tromey
2013-10-05  0:34           ` Doug Evans
2013-10-07 16:31             ` Tom Tromey
2013-09-22  6:25         ` Yao Qi
2013-09-23  0:14           ` Doug Evans
2013-09-24  2:31             ` Yao Qi
2013-10-05  0:37               ` Doug Evans
2013-09-20 19:14       ` Tom Tromey
2013-09-19 17:25   ` [RFC 0/3] GDB Performance testing Doug Evans

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1377663394-4975-1-git-send-email-yao@codesourcery.com \
    --to=yao@codesourcery.com \
    --cc=gdb-patches@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).