townforge/compute
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
1,404 Commits 1 Branch 0 Tags 3.4 MiB
cc
Commit Graph

1404 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Jakub Szuppe
3493b750e8 Fix type, long_ and long might be different types 2016-05-21 16:57:45 +02:00
Jakub Szuppe
bba6c61ed4 Add tests for copying host->device with mismatched types 2016-05-21 16:52:05 +02:00
Jakub Szuppe
a830531445 Fix type-safety issues when copying host->device 2016-05-21 16:44:44 +02:00
Kyle Lutz
d41ff00c06 Merge pull request #613 from haahh/pr_discrete_distribution 
Default constructor for discrete distribution
 2016-05-12 20:23:25 -07:00
Jakub Szuppe
56ad192b92 Now discrete_distribution generates IntType values 
Class discrete_distribution should generate values of type
IntType. Until now it always generated unsigned int values.
 2016-05-11 08:30:46 +02:00
Jakub Szuppe
650e828542 Add more tests for discrete_distribution 2016-05-11 08:30:45 +02:00
Jakub Szuppe
227c05cbdc Add min() and max() methods for discrete_distribution 2016-05-11 08:30:44 +02:00
Jakub Szuppe
e9b2c0ff4b Add default constructor for discrete_distribution 2016-05-11 08:30:41 +02:00
Jakub Szuppe
1799ee1b6e Support for first==last in discrete_distribution range ctor 
Add support for the case with first iterator being equal
to last iterator in range ctor. Now it behaves like ctor in
std::discrete_distribution.
Additionaly: now method probabilities() returns correct vector
of probabilities.
 2016-05-11 08:30:39 +02:00
Kyle Lutz
c173bf02e3 Merge pull request #611 from haahh/pr_coveralls_build 
Dedicated build for coveralls
 2016-05-10 22:06:55 -07:00
Kyle Lutz
d172c8020e Merge pull request #610 from haahh/pr_osx 
OSX
 2016-05-09 21:46:09 -07:00
Jakub Szuppe
476b15933d Remove redundant Travis-CI build 
Remove redundant Travis-CI AMD APP SDK 2.9.1, OpenCL 1.2
build (the Coveralls build has the same test environment,
except for Linux version).
 2016-05-09 12:42:14 +02:00
Jakub Szuppe
a18f7a6dd9 Don't push coverage of tests to the Coveralls 
Coverage info of tests should not be pushed to
the Coveralls service.
 2016-05-09 12:42:14 +02:00
Jakub Szuppe
d57a81d2cd One dedicated build for coveralls 
Instead of enabling coverage analysis in every
build now we have one dedicated build (Trusty,
g++ 4.8, AMD APP SDK v3.0, OpenCL 2.0).
 2016-05-09 12:34:35 +02:00
Jakub Szuppe
13e173c97e Failures not allowed on OSX Travis-CI build 2016-05-05 14:35:21 +02:00
Jakub Szuppe
57ea7d85df Skipping resize_throw_exception test on Apple 
From tests on Travis-CI it seems that on Apple OpenCL platform
clCreatBuffer() does not return NULL and does not set error to
CL_INVALID_BUFFER_SIZE when size of the buffer memory object is
greater than CL_DEVICE_MAX_MEM_ALLOC_SIZE.
 2016-05-05 14:23:28 +02:00
Jakub Szuppe
0fe1ab3e24 Add type checks for distributions 
This commit adds asserts that check if used template
type is correct (is integral, is a floating point type).
 2016-05-05 14:23:28 +02:00
Jakub Szuppe
38b75f920c Fix boost::compute::discrete_distribution constructor 2016-05-05 14:23:28 +02:00
Jakub Szuppe
6412cb2cc6 Fix typos in test_lambda.cpp 2016-05-05 14:23:28 +02:00
Jakub Szuppe
e199304780 Turn off inplace_reduce and radix_sort tests on Apple 
Tests for inplace_reduce, radix_sort and radix_sort_by_key are disabled
for CPU devices on Apple platform because for CPU devices on Apple
platform when local memory is used in a kernel, local work group size
must be [1;1;1]. Those algorithms are designed for GPU devices,
therefore we do no lose any functionality.
 2016-05-05 14:23:28 +02:00
Jakub Szuppe
968056c653 Adapt command queue test to different devices 
Not every device allows for local work group size to be
greater than [1;1;1].
 2016-05-05 14:23:28 +02:00
Jakub Szuppe
af7b197e68 Skip matrix transpose example on Apple CPU devices 2016-05-05 14:21:54 +02:00
Kyle Lutz
dc7bfa0252 Merge pull request #609 from haahh/pr_copy_buffer_for_device_ptr 
Enable copying with copy buffer op for device_ptr
 2016-05-03 20:58:09 -07:00
Kyle Lutz
7b4e652ff7 Merge pull request #608 from haahh/pr_remove_deprecated_macros 
Remove deprecated macros
 2016-05-03 19:33:53 -07:00
Kyle Lutz
8cb06e1954 Merge pull request #607 from jagerman/literal-simplify-test 
Literal test simplification
 2016-05-03 19:31:32 -07:00
Jakub Szuppe
02e12b40fd Enable copying with copy buffer op for device_ptr 
This commit enables copying data using clEnqueueCopyBuffer()
in boost::compute::copy() algorithm for device_ptr<>.
 2016-05-03 21:39:19 +02:00
Jakub Szuppe
2fdafa2128 Remove deprecated Boost macros 2016-05-03 17:13:08 +02:00
Jason Rhinelander
9b2c65fee9 Work around libc++ bug 17782 
libc++ fails to extract a valid float/double literal when it is followed
by any special characters that can also be found in a float value,
in this case, f (https://llvm.org/bugs/show_bug.cgi?id=17782).

This commit works around the problem by not putting the "f" suffix into
the string for extraction.
 2016-05-03 10:21:02 -04:00
Jason Rhinelander
d02b36e735 Simplify literal conversion tests 
The tests are failing under OS X/clang, but it isn't obvious to me why.
This commit simplifies the test code by ignoring the stringstream I/O
status, and reorganizes it to be clearer whether it's the initial
conversion to string that is failing, or the subsequent conversion back
from string.
 2016-05-03 10:21:02 -04:00
Jakub Szuppe
5e28138a7a Update minimum required Boost version to 1.54 2016-05-03 14:32:22 +02:00
Kyle Lutz
51afcd0c22 Merge pull request #602 from boostorg/reduce-uchar-to-float-test 
Add test for reduce() with uchar to float
 2016-05-02 19:44:18 -07:00
Kyle Lutz
90130f8376 Merge pull request #603 from boostorg/fix-sign-comparison-warning 
Fix sign comparison warning in test_literal_conversion
 2016-05-02 19:43:21 -07:00
Kyle Lutz
822be4d306 Merge pull request #605 from boostorg/kernel-get-arg-info-specializations 
Add specializations for kernel::get_arg_info()
 2016-05-02 19:42:13 -07:00
Kyle Lutz
fa6329e8cb Merge pull request #604 from boostorg/fix-function-static-assert-msg 
Fix static assert message in boost::compute::function
 2016-05-02 19:41:58 -07:00
Kyle Lutz
71f1d0d441 Add specializations for kernel::get_arg_info() 2016-04-30 10:32:13 -07:00
Kyle Lutz
52b78417f3 Fix static assert message in boost::compute::function 2016-04-30 09:47:04 -07:00
Kyle Lutz
d15cb3314b Fix sign comparison warning in test_literal_conversion 2016-04-30 09:21:44 -07:00
Kyle Lutz
0b3d80c35a Add test for reduce() with uchar to float 2016-04-29 20:06:52 -07:00
Kyle Lutz
62d487c036 Merge pull request #601 from boostorg/stable-sort-by-key-doc 
Add stable_sort_by_key() to API reference
 2016-04-26 21:36:37 -07:00
Kyle Lutz
1a3cb1468c Add stable_sort_by_key() to API reference 2016-04-26 19:43:51 -07:00
Kyle Lutz
3c2bbbe845 Merge pull request #597 from haahh/pr_fix_find_end 
Fix find_end algorithm
 2016-04-25 21:37:46 -07:00
Kyle Lutz
ecdb24ef86 Merge pull request #595 from haahh/pr_fix_uni_real_dist 
Fix uniform real distribution
 2016-04-25 21:35:23 -07:00
Kyle Lutz
2b346dbd6b Merge pull request #598 from jagerman/increase-literal-precision 
Increase make_literal precision
 2016-04-25 11:59:13 -07:00
Jakub Szuppe
8aea290320 Make sure a < b in uniform real distribution 
Making sure that the left endpoint (a) of the range in uniform
real distribution is less than the right endpoint (b).
 2016-04-25 19:46:21 +02:00
Jakub Szuppe
55bea58790 Add test for values generated by uniform real dist 
New test checks if values generated by uniform real
distribution are in the [a,b) interval.
 2016-04-25 19:46:09 +02:00
Jakub Szuppe
43a2d0286b Fix range of values generated by uniform real dist 
Now generated floating-point values are uniformly distributed
on the interval [a, b). Before this commit it was possible
to produce values equal to b.
 2016-04-25 19:43:47 +02:00
Kyle Lutz
8e5151aec8 Merge pull request #583 from boostorg/fix-intel-platform-documentation 
Fix platform documentation
 2016-04-25 10:01:04 -07:00
Jakub Szuppe
940f379aa9 Fix find_end algorithm 
There was an uninitialized memory at the end of matching_indices
vector and it was read by detail::find_end_helper(...) function
resulting in wrong index returned by this function.
 2016-04-25 13:58:55 +02:00
Jason Rhinelander
7f18293526 Increase make_literal precision 
make_literal is losing some precision when making literal floating
point values because it uses digits10, but that only gives us the
number of decimal digits that will survive a decimal->native>decimal
conversion; what we are doing is a native->decimal->native conversion,
which requires the user of ::max_digits10 (which is 2 (double) or 3
(float) larger than ::digits10).

max_digits10 is a c++11 feature, however, so this commit uses
digits10 + 3 when the c++11 numeric_limits isn't available.
 2016-04-24 17:32:55 -04:00
Jakub Szuppe
b18b5dbe8b Fix test_find_end test 
We have to use the same queue in the whole test.
 2016-04-24 21:46:07 +02:00