Kyle Lutz
417cb03670
Fixes for Mac OS X
2014-12-24 09:25:55 -08:00
Kyle Lutz
88811e345b
Merge pull request #353 from kylelutz/more-thrust-benchmarks
...
Add more Thrust performance benchmarks
2014-12-21 09:34:35 -08:00
Kyle Lutz
e92eb1da54
Add more Thrust performance benchmarks
2014-12-21 09:24:15 -08:00
Kyle Lutz
c97c83a562
Rename types/builtin.hpp to types/fundamental.hpp
2014-12-20 13:39:50 -08:00
Denis Demidov
51db1b1f75
Show Thrust results on performance plots
2014-12-15 23:08:09 +03:00
Denis Demidov
292e5136ed
Use loglog scale for plotting performance data
...
This makes comparison of Boost.Compute and STL performance easier,
especially for smaller sizes.
2014-12-15 16:16:17 +03:00
Denis Demidov
077a4896c7
Make performance scripts compatible with python3
2014-12-15 16:15:37 +03:00
Rastko Anicic
ae014dc757
max_element perf added
2014-12-10 23:23:50 -05:00
Kyle Lutz
05afa5f707
Use thread-local storage for global program cache
2014-09-28 12:37:48 -07:00
Kyle Lutz
f3f1c390de
Add performance section to documentation
2014-09-21 18:36:58 -07:00
Ed Baunton
0be4019ef5
Style
2014-08-25 20:12:58 +02:00
Ed Baunton
0ced318a38
Improve the warning when perf test hasn't been compiled
2014-08-25 20:11:38 +02:00
Kyle Lutz
744359715f
Implement is_sorted() with adjacent_find()
2014-08-21 22:43:46 -07:00
Kyle Lutz
b3ea818248
Rewrite unique() algorithm
2014-08-21 07:00:10 -07:00
Kyle Lutz
ade2ffa067
Fix perf_merge benchmark with odd-sized inputs
2014-08-13 20:53:30 -07:00
roshanr
d356a2c40d
Add benchmark for nth_element
2014-07-30 18:18:34 +05:30
roshanr
5df5d8c3ee
Add benchmark for binary_find
2014-07-18 12:32:18 +05:30
Kyle Lutz
e8afaf1e7c
Add perf_copy_if benchmark
2014-07-12 14:19:23 -07:00
Kyle Lutz
fb5c5442f7
Add perf_fill benchmark
2014-07-12 13:47:01 -07:00
Kyle Lutz
a4ae254adc
Rename mersenne_twister_engine header
...
This renames the mersenne_twister_engine header from
"mersenne_twister.hpp" to "mersenne_twister_engine.hpp".
2014-07-11 22:27:16 -07:00
roshanr
cd41289bbc
Add benchmarks for the new disributions
2014-07-02 20:00:23 +05:30
roshanr
acf4698af1
Add benchmark for lce
...
Added benchmark for lce and changed threads to 1024
2014-06-30 23:41:58 +05:30
roshanr
be18866bca
Add benchmarks for permutation algorithms
2014-06-27 08:22:32 +05:30
roshanr
7baff5980a
Add benchmarks for search algorithms
2014-06-26 01:27:58 +05:30
roshanr
2e825a9ad1
Add benchmarks for partitioning algorithms
2014-06-25 04:43:47 +05:30
roshanr
e2c5fe5504
Add benchmarks for set algorithms
2014-06-21 02:51:00 +05:30
Benoit Dequidt
fed05b3ba5
Update perf_partial_sum : no internal allocation and copy
...
Calling the partial_sum algorithm with the same input/output
buffer leads to an internal allocation and copy.
2014-05-14 23:55:14 +08:00
Benoit Dequidt
48687cd29e
Add perf tests for exclusive_scan
...
- perf_exclusive_scan for Boost.Compute
- perf_thrust_exclusive_scan for thrust
2014-05-14 23:54:58 +08:00
Kyle Lutz
c2ad69aaa0
Fix BOOST_COMPUTE_FUNCTION() usage in perf_cart_to_polar
2014-04-22 21:33:44 -07:00
Kyle Lutz
2511bdb436
Merge pull request #97 from roshanr95/unique
...
Fix errors in unique
2014-04-20 16:42:56 -07:00
roshanr
59efaa0eeb
Add tests and benchmarks for unique_copy
2014-04-21 02:00:02 +05:30
Kyle Lutz
8b06e3f7bb
Add event::duration() method
2014-04-19 12:31:37 -07:00
roshanr
f491c55dbd
Change perf_sort_by_key to use int_ and long_ instead of int and long
2014-04-18 12:17:55 +05:30
Kyle Lutz
6336b81911
Rename mersenne_twister_engine::fill() to generate()
2014-04-13 14:13:53 -07:00
Kyle Lutz
fa8fcade52
Add perf_erase_remove benchmark
2014-04-12 16:07:30 -07:00
Kyle Lutz
b7c4f0ce18
Change mersenne_twister::seed() to take a command_queue
2014-04-12 11:14:44 -07:00
roshanr
fca6e136c2
Fix unique benchmarks
2014-04-03 12:07:16 +05:30
roshanr
1e81b7ec2e
Unique algorithm
...
Added unique() algorithm, tests and benchmarks. Removed unused variable
in scan_on_gpu() to remove warnings
2014-03-24 06:30:28 +05:30
roshanr
9450014389
Add benchmarks for rotate and rotate_copy
...
Added compute and stl benchmarks for rotate() and rotate_copy()
2014-03-17 19:12:49 +05:30
Kyle Lutz
da22127181
Merge pull request #67 from roshanr95/reverse_benchmarks
...
Add benchmarks for reverse
2014-03-16 10:27:14 -07:00
roshanr
a65804d0c9
Add benchmarks for reverse
...
Added STL and Compute benchmarks for reverse()
2014-03-16 22:31:29 +05:30
Kyle Lutz
bae7432c04
Improve sort_by_key() performance
2014-03-12 23:40:57 -07:00
Kyle Lutz
9e5bff9d83
Add SYSTEM to third-party include_directories() calls
2014-03-12 20:41:04 -07:00
Kyle Lutz
ec11d8cdc4
Add third-party perf tests
...
This adds third-party performance tests to use in comparing
Boost.Compute with other parallel/GPGPU frameworks like Intel's
TBB and NVIDIA's Thrust along with the C++ STL.
Also refactors the timing and profiling infrastructure and adds
a simple perf.py driver script for running performance tests.
2014-02-02 13:12:17 -08:00
Kyle Lutz
ccd6f21d98
Change vector constructors to take queue argument
...
This changes the vector<T> constructors which copy or initialize
data to take a queue argument used for performing the operations.
Previously they just took a context argument used to initialize the
buffer and then created a new command queue to use. This improves
performance by not requiring a new command queue and also fixes issues
when performing operations on a different command queue while the
vector was still being initialized.
2014-01-27 23:39:19 -08:00
Denis Demidov
562f149b18
Implements offline kernel caching
...
See kylelutz/compute#21
This adds program::build_with_source() function that both creates and
builds the program for the given context with supplied source and
compile options. In case BOOST_COMPUTE_USE_OFFLINE_CACHE macro is
defined, it also saves the compiled program binary for reuse in the
offline cache located in $HOME/.boost_compute folder on UNIX-like
systems and in %APPDATA%/boost_compute folder on Windows.
All internal uses of program::create_with_source() followed by
program::build() are replaced with program::build_with_source().
2014-01-07 09:07:00 +04:00
Kyle Lutz
84b45b4394
Bump Boost version requirement to 1.48
...
This bumps the required Boost version to 1.48 as Boost.Compute
makes use of Boost.Move which was added in version 1.48.
See issue #37 .
2014-01-05 18:53:19 -08:00
Kyle Lutz
8b812a2081
Add perf_cart_to_polar benchmark
...
This adds a new benchmark to measure the performance of converting
cartesian to polar coordinates using the transform() algorithm.
2013-11-23 12:26:10 -08:00
Kyle Lutz
5b7e22bac2
Add perf_mersenne_twister benchmark
...
This adds a benchmark for the mersenne_twister engine.
2013-09-24 22:51:26 -04:00
Kyle Lutz
07e4a6b3aa
Remove BLAS functions
...
This removes the incomplete BLAS API functions.
2013-09-24 22:19:56 -04:00
