Commit Graph

383 Commits

Author SHA1 Message Date
Eric Fiselier
279e502a05 add test for benchmark with two template parameters 2015-03-17 13:32:20 -04:00
Eric Fiselier
efe591015d address review comments 2015-03-17 12:31:46 -04:00
Eric Fiselier
66bf7c8f71 add floating point comparison warnings 2015-03-12 20:27:29 -04:00
Eric Fiselier
7a767012f1 Adopt new benchmark timing internals.
This patch adopts a new internal structure for how timings are performed.
Currently every iteration of a benchmark checks to see if it has been running
for an appropriate amount of time. Checking the clock introduces noise into
the timings and this can cause inconsistent output from each benchmark.

Now every iteration of a benchmark only checks an iteration count to see if it
should stop running. The iteration count is determined before hand by testing
the benchmark on a series of increasing iteration counts until a suitable count
is found. This increases the amount of time it takes to run the actual benchmarks
but it also greatly increases the accuracy of the results.

This patch introduces some breaking changes. The notable breaking changes are:
1. Benchmarks run on multiple threads no generate a report per thread. Instead
   only a single report is generated.
2. ::benchmark::UseRealTime() was removed and replaced with State::UseRealTime().
2015-03-12 18:03:33 -04:00
Dominic Hamon
d68127d8ad Enable UseRealTime and fix documentation for SetLabel.
Fixes #89

UseRealTime was defined in the internal namespace by mistake.
Similarly, documentation suggested that benchmark::SetLabel should be
used to set a label, and a function was declared but not defined, while
actually the call should be 'state.SetLabel'.
2015-03-11 09:47:15 -07:00
Eric Fiselier
e975efdb7a remove third_party directory. update test args 2015-03-10 14:38:45 -04:00
Dominic Hamon
3b40f0a7a3 Add filter test, remove re test, and googletest deps 2015-03-09 20:30:14 -07:00
Eric Fiselier
a187aa0803 Third step in moving towards adopting new timer semantics 2015-03-06 17:01:05 -05:00
Eric Fiselier
f947cebe15 remove ifdefs around BM_Factorial 2015-03-06 16:59:32 -05:00
Eric Fiselier
8f96f50553 comment out test to prevent failures. 2015-03-06 16:53:52 -05:00
Eric Fiselier
8ed7d7664b The second step towards merging the timer changes.
This patch does two things:

1. It overhalls the static initialization in Walltime to be simplier. It uses
   a static variable inside WallTime::Now() to initialize the timer.

2. Add a logging mechanism so that the -v flag actually has meaning and
   reimplement the CHECK macros to allow extra messages to be streamed in.
2015-03-06 14:07:24 -05:00
Eric Fiselier
b1f34d3880 address comments 2015-03-06 13:20:47 -05:00
Eric Fiselier
5b41e128b3 Step one towards merging timer changes.
This patch cleans up our use of generic macros and also merges changes in the
build system.

It adds options -DBENCHMARK_ENABLE_TESTING and -DBENCHMARK_ENABLE_SHARED.
2015-03-06 12:35:00 -05:00
Dominic Hamon
3968ff45ba Fix #72 by avoiding 64-to-32-bit shortenings 2015-02-18 11:18:40 -08:00
Dominic Hamon
b40b66ab3b Actually fix issue with CHECK 2014-12-02 12:58:14 -08:00
Dominic Hamon
d31977bb69 Added gtest CHECK symbol to benchmark test 2014-12-02 12:51:46 -08:00
Lei Xu
0f78cebba5 Use #ifdef to protect calling Factorial in benchmark_test.cc 2014-11-13 16:09:21 -08:00
Dominic Hamon
7e290fb965 Fix release builds 2014-11-13 13:12:28 -08:00
Dominic Hamon
a3b5e44c52 Remove CHECK_* from public API.
CHECK_* are now private and used internally in the library. The test
uses have been replaced with asserts.

Fixes #62.
2014-10-31 21:54:55 -07:00
Niklas Hofmann
b04211cc88 Merge branch 'master' of github.com:google/benchmark 2014-10-25 18:35:35 +02:00
Niklas Hofmann
3e9264dcf1 added googletest dependency to tests (allows compilation with -j) 2014-10-16 19:17:08 +02:00
Dominic Hamon
a13627bc3a Move pthread dependency to test CMakeLists.txt 2014-10-11 14:57:45 -07:00
Matt Clarkson
0fd0be9789 Added more complicated regex test patterns 2014-08-22 14:55:46 +01:00
Matt Clarkson
1176936966 Check the number of benchmark tests ran
Previously the benchmark_test program executed the benchmark tests to make sure
the API was working but was not checking the number of tests that were
completed. If the regular expression matching breaks, zero tests could be ran.
Similarly, if one of the APIs breaks and doesn't run the correct amount of tests
then `make test` will catch this.
2014-08-15 11:09:23 +01:00
Matt Clarkson
373cc41100 C++11 concurrency instead of pthread 2014-08-06 10:50:21 -07:00
Paul Redmond
52b3004ae3 fix linker error by reordering link libraries 2014-07-23 10:35:42 -04:00
Chris Kennelly
fa908926c7 Partially resolve google/benchmark#17 by fixing regular expression leak.
This adds a unit test to validate the wrapper without running the entirety of
benchmark_test.
2014-04-23 00:56:17 -07:00
Chris Kennelly
e38fde6450 Setup targets from each directory, rather than the top-level. 2014-04-23 00:47:07 -07:00
Shuo Chen
664159372c Use a new container in each round of push_back test.
Otherwise containers keep growing and uses gigabytes of memory.
2014-04-17 10:04:02 -07:00
Dominic Hamon
3a01f8320b Add items/bytes processed to SetInsert benchmark test 2014-01-07 14:31:14 -08:00
Dominic Hamon
2923a481ed Multithreaded tests are reenabled 2014-01-07 13:15:26 -08:00
Dominic Hamon
15bf66750c Really drop multithreading support 2013-12-20 15:10:46 -08:00
Dominic Hamon
403f354423 Initial commit
Benchmark library builds and runs but only single-threaded. Multithreaded
support needs a bit more love.

Currently requires some C++11 support (g++ 4.6.3 seems to work).
2013-12-18 16:58:23 -08:00