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

791 Commits

Author SHA1 Message Date
Jakub Szuppe
c47753c4fa Stable comparison sort, sort-by-key 2016-06-22 17:06:12 +02:00
Jakub Szuppe
7996382178 Use index instead of copying vals in bitonic sort-by-key 
Instead of copying values in local memory in bitonic block
sort (when sorting by key), we save and copy local index
and copy value once at the end. This saves local memory.
 2016-06-22 15:59:55 +02:00
Jakub Szuppe
8f033e72ee Add comparison sort and sort by key 2016-06-21 17:41:36 +02:00
Jakub Szuppe
8f4db3d7b7 Reuse dispatch_copy_async() in dispatch_copy() 2016-05-28 22:54:51 +02:00
Jakub Szuppe
bffc9f74fb Fix async copying when first == last (input is empty) 2016-05-28 22:46:35 +02:00
Jakub Szuppe
c9a0aba172 Remove debug macro, fix names of event variables 2016-05-28 22:05:50 +02:00
Jakub Szuppe
5c29b50823 Coping SVM to/from host by mapping SVM memory 2016-05-28 13:15:36 +02:00
Jakub Szuppe
753f883321 Support for svm_ptr<T> in meta_kernel 2016-05-28 13:15:36 +02:00
Jakub Szuppe
4280024f26 Now svm_ptr<T> keeps its context 
This commit modifies svm_ptr<T> to keep its context. It is
convenient for the users and enables creating
svm_ptr_index_expr<T, IndexExpr> class.
 2016-05-28 13:15:36 +02:00
Jakub Szuppe
aaaaca4ed2 Type-safe async coping from device to host 2016-05-28 13:15:36 +02:00
Jakub Szuppe
3dcbd4c001 Fix async coping svm_ptr<> from/to/on device 2016-05-28 13:15:36 +02:00
Jakub Szuppe
1419b7d36f Type-safe copying from device to host 
Type-safe copying from device to host. Seperate copying
algorithm device -> host for non-contiguous OutputIterator (host).
 2016-05-24 21:37:29 +02:00
Jakub Szuppe
ce8230b6f5 Coping from host for non-contiguous iterators (host) 
When InputIterator (host) is a non-contiguous iterator we don't
need a separate algorithm for cases when value_types of InputIterator
and OutputIterator (device) do not match and cases when they do
match.
 2016-05-21 17:16:13 +02:00
Jakub Szuppe
d12e07c0e9 Fix type-safety issues for async copying host->device 2016-05-21 16:58:15 +02:00
Jakub Szuppe
a830531445 Fix type-safety issues when copying host->device 2016-05-21 16:44:44 +02:00
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
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
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
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
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
Jakub Szuppe
5e28138a7a Update minimum required Boost version to 1.54 2016-05-03 14:32:22 +02:00
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
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
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
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
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
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
6d5d9b0b5d Now radix_sort_by_key is stable 
Radix sort algorithm was not stable when used for
sorting by key in descending order (see issue #588).
 2016-04-23 18:21:15 +02:00
Jason Rhinelander
d4653c6624 Use nextafter to push values down into [0,1) range 
Without this, we can get values of 1, but that leads to negative
infinity from the following log; if we push the x1 value towards 0, we
ensure that (one-x1) is always a value in (0,1], and so the log value
will be in (-inf, 0], which is what we want.
 2016-04-22 19:27:47 -04:00
Kyle Lutz
a1090c1290 Fix issue with zero values in normal_distribution 2016-04-22 18:51:26 -04:00
Kyle Lutz
bf8eeb9738 Use natural log in normal_distribution kernel 2016-04-22 18:50:49 -04:00
Kyle Lutz
3d9fd19350 Merge pull request #585 from bwitherspoon/wait-list-initializer-list 
Add initializer list constructor to wait_list
 2016-04-18 14:51:16 -07:00
Brett Witherspoon
8f449cde4b Add initializer list constructor to wait_list 2016-04-17 02:59:51 -05:00
Jakub Szuppe
da222c8243 Remove unused variable 2016-04-10 12:46:38 +02:00
Kyle Lutz
44f08e3932 Merge pull request #580 from boostorg/fix-svm-ptr-assignment-operator 
Add missing return to svm_ptr assignment operator
 2016-04-07 15:10:42 -07:00
Kyle Lutz
203f0be723 Add missing return to svm_ptr assignment operator 2016-04-07 07:58:03 -07:00
Kyle Lutz
523d8e9745 Merge pull request #577 from ddemidov/wait_list-improvements 
Wait list improvements
 2016-03-18 19:38:41 -07:00
Kyle Lutz
da71051593 Merge pull request #573 from haahh/pr_stable_sort_by_key 
Add stable_sort_by_key()
 2016-03-18 19:36:22 -07:00
Denis Demidov
cafe8b6578 Provide wait-list iterators 2016-03-18 22:27:55 +03:00
Denis Demidov
15b7414e78 Provide array subscript operator for wait_list 2016-03-18 22:20:39 +03:00
Denis Demidov
2eed2c2fdd Provide wait_list::reserve() method 
This should help to get rid on unnecessary heap allocations.

The wait lists are usually small, which means that just pushing back
elements without prior call to reserve has very high chance of hitting
capacity limit every time wait_list::insert is called.
 2016-03-18 10:58:29 +03:00
Jakub Szuppe
73ef474fdd Add stable_sort_by_key() 
Currently, stable_sort_by_key() works on GPUs only for
keys that are radix-sortable and for very small inputs.
 2016-03-16 14:59:02 +01:00
Jakub Szuppe
b3224ea753 Misc. strided_iterator changes 2016-03-16 13:16:14 +01:00
Jakub Szuppe
b8fa408d4b Fix compilation errors on OSX 
Fix compilation errors occurring on OSX in strided_iterator,
merge_with_merge_path() and includes(). They were introduced
by commit 3826e9cae4 fixing
signed/unsigned int comparison warnings on Windows.
 2016-03-16 13:08:39 +01:00
Kyle Lutz
38ba1c5447 Merge pull request #566 from haahh/update-test-jamfilev2 
Update tests compile options in Jamfile.v2
 2016-03-10 13:04:02 -08:00
Kyle Lutz
fe816c5acf Merge pull request #571 from haahh/win-related-changes 
Windows related changes
 2016-03-10 10:14:03 -08:00
Jakub Szuppe
3826e9cae4 Misc fixes 2016-03-09 18:56:17 +01:00
Jakub Szuppe
1653a06f90 Suppress OpenCL-related deprecated declarations warns 
Suppress deprecated declarations warnings (for MSVC, GCC and Clang)
caused by deprecated OpenCL Runtime functions.
 2016-03-09 17:08:45 +01:00
Kyle Lutz
03dc34b820 Add type_name() support for OpenGL types 2016-03-07 19:52:32 -08:00
Jakub Szuppe
c7b87432cb Fix valarray operators 2016-02-16 21:28:25 +01:00
Kyle Lutz
ff0b5ad69e Add missing forward declaration to image3d.hpp 2016-02-02 20:45:10 -08:00
Kyle Lutz
9f1766dd17 Fix counting_iterator with unsigned types 2016-01-18 20:53:41 -08:00
Kyle Lutz
3466881454 Merge pull request #548 from haahh/timberosx_fixes 
Timber OSX fixes
 2016-01-13 10:56:58 -08:00
Jakub Szuppe
6b34c63fb8 Fix kernels in reduce_by_key_with_scane.hpp 
It was not guaranteed that every barrier would be encountered by all
work-items in a work-group executing the kernels due to some work-items
returing too early, not entering the conditional if with with a barrier
or not executing every iteration of the loop with a barrier.

See barrier function description in OpenCL documentation.
 2016-01-07 20:55:00 +01:00
Jakub Szuppe
6396530a2b Fix comparison between signed and unsigned integer 2016-01-07 20:54:54 +01:00
Jakub Szuppe
a6f258cf09 Fix find_extrema_reduce kernel in find_extrema_with_reduce.hpp 
It was not guaranteed that every barrier would be encountered by all
work-items in a work-group executing the kernel due to some work-items
returing too early, not entering the conditional if with with a barrier
or not executing every iteration of the loop with a barrier.

See barrier function description in OpenCL documentation.
 2016-01-07 20:51:56 +01:00
Kyle Lutz
767c5beba8 Merge pull request #545 from okdshin/fix_linear_congruential_rng_copy 
Modify linear_congruential_engine to copy its all states.
 2015-12-28 12:28:58 -08:00
okdshin
5c467c9850 modify linear_congruential rng to copy its all states 2015-12-29 00:51:14 +09:00
okdshin
babfbafad6 Modify mersenne_twister_engine to copy its states. 2015-12-27 04:24:55 +09:00
Kyle Lutz
301c541824 Release v0.5 2015-12-20 18:34:14 -08:00
Kyle Lutz
0d846daad9 Merge pull request #542 from haahh/count_fix_rngs 
Fix incorrect count update in rngs
 2015-12-19 09:05:44 -08:00
Jakub Szuppe
39fa3634e1 Fix seed method call in linear_congruential_engine.hpp 
There is no seed method with just one parameter of type result_type.
 2015-12-18 22:12:51 +01:00
Jakub Szuppe
d2d62f9973 Fix incorrect count (global work size) update in generate methods 2015-12-18 22:12:47 +01:00
Jakub Szuppe
1f4088200d Minor fixes related to Boost Inspection Report (1.60 develop) 2015-12-06 18:35:29 +01:00
Kyle Lutz
cb863b91a6 Add program::create_with_source() with vector of strings 2015-12-04 20:37:13 -08:00
Kyle Lutz
c983cc3a48 Merge pull request #535 from olivas/create_with_source_file 
program::create_with_source_file doesn't fail with bad input
 2015-11-19 19:39:58 -08:00
Alex Olivas
0a2d7fa0e1 if the ifstream can't be created in program::create_with_source_file then it throws std::ios_base::failure. 2015-11-18 16:27:19 -07:00
Kyle Lutz
0d172afaf8 Merge pull request #533 from haahh/minor_fixes 
Minor fixes
 2015-11-05 18:53:42 -08:00
Jakub Szuppe
6be8228a74 Fix integer type to uint_ (cl_uint) 2015-11-05 18:23:00 +01:00
Jakub Szuppe
e97f70cdac CL_MAP_WRITE_INVALIDATE_REGION is available since OpenCL 1.2 2015-11-05 18:15:39 +01:00
Kyle Lutz
f249741bf8 Merge pull request #522 from haahh/pr_find_extrema_first_extremum 
Fix find_extrema_with_atomics and _with_reduce
 2015-10-23 00:19:54 -07:00
Jakub Szuppe
8099e07f40 Fix find_extrema_with_atomics 
Now find_extrema_with_atomics always return the first extremum
in the given range.
 2015-10-21 08:46:04 +02:00
Jakub Szuppe
2d972fe2a4 Fix find_extrema_with_reduce 
Now find_extrema_with_reduce always return the first extremum
in the given range.
 2015-10-21 08:46:04 +02:00
Kyle Lutz
781cf302cb Merge pull request #525 from haahh/pr_vector_custom_alloc 
Vector with custom allocator fix
 2015-10-10 09:16:11 -07:00
Jakub Szuppe
afd09bcd7d Fix vector when used with custom allocator 2015-10-05 14:28:37 +02:00
Kyle Lutz
0b4509afea Merge pull request #520 from jpola/scatter_if 
scatter if operation
 2015-10-01 20:40:39 -07:00
Kyle Lutz
e27e5ee72b Merge pull request #518 from boostorg/event-wait-const 
Event wait const
 2015-09-28 19:50:29 -07:00
Kyle Lutz
b2e2bdfb0d Merge pull request #510 from haahh/dev_events_for_queue_enqueue_methods 
Events for queue enqueue_map_ methods
 2015-09-28 19:50:10 -07:00
Jakub Pola
c6c7f4a679 Copyrights fixes 2015-09-28 21:16:31 +02:00
Jakub Pola
f25b47d6f3 scatter if operation 2015-09-28 17:17:10 +02:00
Kyle Lutz
a56ae5687f Make wait_list::wait() const 2015-09-23 21:39:47 -07:00
Kyle Lutz
c248dadbca Make event::wait() const 2015-09-23 21:39:02 -07:00
Kyle Lutz
c75f5dc858 Merge pull request #516 from boostorg/adapt-struct-array 
Add array support to BOOST_COMPUTE_ADAPT_STRUCT()
 2015-09-20 16:58:55 -07:00
Kyle Lutz
68155f7597 Add array support to BOOST_COMPUTE_ADAPT_STRUCT() 
This adds support for C-style arrays (e.g. "int array[10]") to the
BOOST_COMPUTE_ADAPT_STRUCT() macro.

Thanks to Fabian Bösch for providing the code.
 2015-09-19 18:28:17 -05:00
Jakub Szuppe
36418e4cf8 Remove redundant kernel compilation from binary_find 
After this binary_find kernel is compiled only once during binary_find algorithm
execution instead of in every iteration of the while loop.
 2015-09-19 19:17:27 +02:00
Jakub Szuppe
10583d3868 Fix binary_find algorithm 
This prevents iterators from going out of range which
resulted in errors.
 2015-09-19 19:11:32 +02:00
Kyle Lutz
f04b1aabde Merge pull request #511 from haahh/pr_fix_search_search_n 
Fix for search and search_n algorithm
 2015-09-15 12:41:10 -05:00
Jakub Szuppe
0e51ad4ad8 Fix for search and search_n algorithm 2015-09-12 19:29:16 +02:00
Jakub Szuppe
55a7f874a6 Add missing enqueue_map_image method in command queue 2015-09-12 16:26:10 +02:00
Jakub Szuppe
88beaa0502 Add asynchronous versions of a few methods in command_queue class 
Add asynchronous versions for enqueue_map_buffer(), enqueue_read_buffer_rect()
and enqueue_write_buffer_rect().
 2015-09-12 16:26:09 +02:00
Jakub Szuppe
5c99e3f7f3 enqueue_map_buffer returns associated event through parameter 2015-09-12 16:26:09 +02:00
Kyle Lutz
75d32254b1 Merge pull request #509 from haahh/pr_algorithm_reduce_by_key 
Include Reduce by key in algorithm.hpp meta-header
 2015-09-11 19:17:52 -07:00
Jakub Szuppe
1566a79a4e Add reduce by key algorithm to algorithm.hpp meta-header 2015-09-11 21:51:01 +02:00
Lorenzo Pistone
9211260bcc fix a few instances where the device type is not used as a bit mask 2015-09-11 12:21:11 +02:00
Kyle Lutz
239f4a247c Merge pull request #498 from boostorg/invoke 
Add invoke() utility function
 2015-09-01 21:49:54 -07:00
Kyle Lutz
5ac31a7f5f Merge pull request #496 from haahh/pr_enqueue_methods_return_events 
Return events from all command_queue::enqueue_* methods
 2015-09-01 21:49:23 -07:00
Jakub Szuppe
c882466dbd enqueue_* methods in command_queue class now return event object 
All enqueue_* methods in command_queue class (except for
enqueue_map_buffer() and enqueue_barrier()) now return event object.
 2015-08-20 12:01:55 +02:00
Kyle Lutz
cf5e40ee26 Merge pull request #497 from haahh/pr_find_extrema_supports_custom_comparision 
Support for custom comparison function in find_extrema()
 2015-08-19 20:24:04 -07:00
Kyle Lutz
ef109b75d5 Add invoke() utility function 
This adds an invoke() function which calls an OpenCL function on a
device with the given arguments.
 2015-08-19 08:18:15 -07:00
Jakub Szuppe
e6b20e6af7 Rename find_extrema_reduce() to find_extrema_with_reduce() 2015-08-19 13:42:22 +02:00
Jakub Szuppe
a50dae747d Support for custom comparision function in find_extrema() and min/max_element() 2015-08-19 13:42:22 +02:00
Kyle Lutz
057b220d18 Merge pull request #495 from haahh/pr_improve_find_if 
Checking multiple values per thread in find_if()
 2015-08-18 08:27:20 -07:00
Jakub Szuppe
75a42646a1 Checking multiple values per thread in find_if_with_atomics() 2015-08-15 15:39:06 +02:00
Kyle Lutz
fd43795770 Merge pull request #491 from haahh/pr_always_serial_reduce_for_cpus 
Using serial_reduce() when reducing on CPU device
 2015-08-12 08:30:14 -07:00
Jakub Szuppe
d98e6dfee6 Using serial_reduce() when reducing on CPU device 2015-08-08 18:00:49 +02:00
Jakub Szuppe
de0c8ba177 CPU-targeted merge sort adapted for sorting by key 2015-08-03 23:37:48 +02:00
Kyle Lutz
cadff8e049 Merge pull request #489 from haahh/pr_serial_merge 
Using serial merge in merge() for small inputs
 2015-08-01 13:44:52 -07:00
Jakub Szuppe
e3324f887d Using serial merge in merge algorithm for small inputs 2015-08-01 19:41:16 +02:00
Kyle Lutz
f278c66c14 Merge pull request #486 from haahh/pr_sort_cpu 
CPU-targeted merge sort
 2015-08-01 09:10:13 -07:00
Jakub Szuppe
b6c2e71ba3 Using merge with merge path algorithm in merge sort 
Merge with merge path is used in merge part of merge sort
algorithm as it's seems to be more efficient than block merge
when there are a few big sorted blocks left to be merged.
 2015-07-30 20:56:40 +02:00
Jakub Szuppe
b5bef712cd Merge sort for CPU devices 2015-07-30 20:55:49 +02:00
Kyle Lutz
15a778d84a Merge pull request #480 from haahh/pr_boost_inspection 
Fixing problems reported in Boost Inspection Report
 2015-07-18 11:01:57 -07:00
Jakub Szuppe
7a8d078926 Update Boost macros, macros style unification 2015-07-18 16:12:24 +02:00
Jakub Szuppe
7e12f63a81 Add missing Boost license info and copyrights 2015-07-18 14:01:29 +02:00
Jakub Szuppe
350f2c93a8 Fix violations of Boost min/max guidelines 2015-07-16 21:18:08 +02:00
Jakub Szuppe
c5409541ac Parallel reduce by key algorithm implementation 2015-07-15 11:20:49 +02:00
Jakub Szuppe
706af4fce8 Reduce by key algorithm (serial implementation) 2015-07-15 11:06:23 +02:00
Kyle Lutz
377e509acd Merge pull request #472 from msuchard/master 
Apple NVIDIA vendor name does not include Corporation
 2015-06-24 21:54:23 -07:00
Marc Suchard
7009451dbf Apple NVIDIA vendor does not include Corporation 2015-06-24 14:53:59 -07:00
Kyle Lutz
2298aa5ff1 Merge pull request #468 from haahh/custom_funcs_for_scan 
Support for custom functions in inclusive and exclusive scan
 2015-06-24 08:30:06 -07:00
Jakub Szuppe
7bea0b13a1 Support for a custom function in exclusive and inclusive scan 2015-06-24 08:39:45 +02:00
Kyle Lutz
20ff5cbfb7 Fix compilation error in find_extrema_reduce() 2015-06-23 22:03:19 -07:00
Jakub Szuppe
feeab7a277 Add single-argument constructor for vector types 2015-06-23 10:14:50 +02:00
Kyle Lutz
1c68d8a73b Merge pull request #443 from bkchr/runtime_version_checking 
Runtime version checking
 2015-06-22 19:59:53 -07:00
Bastian Köcher
60f4b31bdd Add runtime version checking in the command_queue functions and fixed bug with context::get_device(). 
This adds runtime version checking in the command_queue functions which fixes issues if you compile
with OpenCL2.x and run your code on OpenCL1.x devices.
Fixed bug with context::get_device() if the context contains more than one device.
 2015-06-04 11:59:54 +02:00
Junaid Muzammil
a9c373f92c Merge remote-tracking branch 'newrepo/develop' into develop 2015-06-02 19:18:00 +05:00
Jakub Szuppe
4c20b3fb52 Add find_extrema based on parallel reduction 
This adds find_extrema_reduce function based on parallel reduction.
Now find_extrema uses find_extrema_with_atomics only if local memory size
requirements for running find_extrema_reduce are not met.
 2015-05-21 13:27:06 +02:00
Kyle Lutz
fda67a22d0 Update GitHub links 2015-05-17 20:32:09 -07:00
Kyle Lutz
ca0bf77551 Fix default values in uniform_int_distribution constructor 
This fixes the default values for the uniform_int_distribution
constructor to match std::uniform_int_distribution.
 2015-05-17 11:37:54 -06:00
Kyle Lutz
4c00484e04 Merge pull request #454 from haahh/pr_valarray_operators 
Missing valarray operators
 2015-05-15 10:41:39 -06:00
Jakub Szuppe
c2c8eb6e54 Fix valarray support for vector types 2015-05-15 18:12:12 +02:00
Jakub Szuppe
332a412d1a Valarray operators work only with fundamental OpenCL types 
Valarray arithmetic, logical and comparison operators work only
when valarray type is built-in OpenCL scalar or vector type.
This also fixes empty macro argument error when compiled with
pedantic and Werror options, and ambiguous 'bind' function call
error.
 2015-05-15 18:12:12 +02:00
Jakub Szuppe
28f08dda44 Add valarray operators 
This adds missing operators to valarray. Note that in OpenCL it's impossible
to create memory buffer with bool type and because of that valarray<char> is used
as a return type for comparison and logical operators instead of valarray<bool>.
Implemented operators work only for default context (queue) and are asynchronous.
 2015-05-15 18:12:07 +02:00
Jakub Szuppe
29dc53b110 Add missing bitwise shift operators 2015-05-15 18:05:44 +02:00
Kyle Lutz
79aa8f9086 Merge pull request #452 from haahh/strided_iterator 
Add strided_iterator class (skip_iterator) #121
 2015-04-20 19:58:31 -07:00
Jakub Szuppe
1ba5c5d950 Fix make_strided_iterator_end() description 
Fix make_strided_iterator_end() description and some comments.
 2015-04-18 10:55:32 +02:00
Jakub Szuppe
6c22df7077 Better descriptions and minor refactoring 2015-04-12 22:45:39 +02:00
Jakub Szuppe
7ce179d414 Add make_strided_iterator_end function 2015-04-12 16:05:22 +02:00
Jakub Szuppe
c80d42aa29 Strided iterator 
Iterator adaptor which skips over multiple elements each time it is incremented.
 2015-04-12 15:56:10 +02:00
Denis Demidov
d4651113e7 Get rid of (potentially huge) temporary strings in offline cache 
See #440, #441
 2015-04-08 15:20:02 +03:00
Anees Haider
b979ff527d Added Threefry RNG to performance benchmark 2015-04-06 09:51:53 +05:00
Anees Haider
b565232b8e Merge remote-tracking branch 'compute/develop' 2015-04-06 09:33:52 +05:00
Junaid Muzammil
31342d7e53 Threefry generate signature modification 2015-04-05 21:27:31 +05:00
Denis Demidov
56ca5f2331 Add opencl vector types to the list of fundamental types 
refs #449
 2015-04-05 15:27:34 +03:00
Kyle Lutz
6a6a2aaf64 Merge pull request #448 from ddemidov/develop 
kernel::get_work_group_info() should be const
 2015-04-04 15:55:19 -07:00
Denis Demidov
a754fcb720 kernel::get_work_group_info() should be const 2015-04-04 22:02:42 +03:00
Muhammad Junaid Muzammil
61e16c8963 Overloaded Method for default Key Values 2015-04-02 00:26:06 +05:00
Kyle Lutz
f9136ed842 Add transform_if() algorithm 2015-03-31 19:45:58 -07:00
Junaid Muzammil
1c41612c9a Unit Test Fixed and Comments Added 2015-03-29 00:57:56 +05:00
Junaid Muzammil
1c75d5752d Necessary Changes Made to Threefry Engine as recommended 2015-03-28 18:06:01 +05:00
Junaid Muzammil
44c54d4f7a Refactored Threefry RNG 2015-03-22 18:04:01 +05:00
Muhammad Junaid Muzammil
37e060d0ca Threefry Random 123 Support Added 2015-03-22 16:53:47 +05:00
Kyle Lutz
34e042a298 Reduce Boost header includes 2015-03-21 11:38:30 -07:00
Kyle Lutz
78db09bf10 Use gcc-4.8 on Travis-CI 2015-03-16 19:55:11 -07:00
Kyle Lutz
4175a85171 Add parameter_cache class 
This adds a parameter_cache class which can be used to store
execution parameters for an algorithm. Also updates some of
the benchmark programs to find and store optimal parameters.
 2015-03-15 18:15:57 -07:00
Jakub Szuppe
f95b1ee682 Separate kernel for reverse_copy() 
Separate kernel for reverse_copy() algorithm improves its performance,
new tests for reverse_copy_int
 2015-03-10 22:25:38 +01:00
Kyle Lutz
d4adfc5feb Add support for OpenCL-OpenCL sharing on Mac OS X 2015-02-26 14:57:26 -07:00
Kyle Lutz
a71a4e9853 Add begin() and end() methods to extents<N> 2015-02-25 11:06:45 -07:00
Shehzan Mohammed
4961a6177f BUGFIX enable double for key type in radix sort by key 2015-02-23 12:15:40 -05:00
Kyle Lutz
de76e66dd1 Merge pull request #426 from jmr1/msvc2013 
MSVC 2013 build fixes
 2015-02-22 19:02:45 -07:00
jmr1
9c3d2e0a87 MSVC 2013 workaround for passing one parameter to a variadic template 2015-02-22 17:33:35 +01:00
Kyle Lutz
ca0bc4b7d3 Refactor image classes 2015-02-08 16:26:34 -08:00
Kyle Lutz
6e38dafbd5 Merge pull request #423 from kylelutz/buffer-clone-flags 
Preserve memory flags in buffer::clone()
 2015-02-08 15:06:57 -08:00
Kyle Lutz
a34aaf0456 Merge pull request #422 from kylelutz/no-device-found-exception 
Add no_device_found exception
 2015-02-08 15:06:43 -08:00
Kyle Lutz
8b2788f026 Add no_device_found exception 2015-02-07 12:47:11 -08:00
Kyle Lutz
3c56ea4cbe Preserve memory flags in buffer::clone() 2015-02-07 12:27:24 -08:00
Kyle Lutz
155ec4b8b9 Disable default context error handler 2015-02-07 10:56:34 -08:00
Bastian Köcher
426fbbb8c4 - Added missing header 2015-01-21 20:03:54 +01:00
Bastian Köcher
1e711e0330 - Rebased with develop branch 
- Moved include into the test

Conflicts:
	test/test_buffer.cpp
 2015-01-20 10:19:37 +01:00
Kyle Lutz
f4bf942a31 Merge pull request #412 from bastiankoe/createsubbuffer 
CreateSubbuffer support
 2015-01-16 18:40:01 -08:00
Bastian Köcher
09d3c44cd7 - Added templated register_destructor_callback function and a test for this new functionality 2015-01-16 16:38:25 +01:00
Bastian Köcher
7eba538c55 Merge branch 'develop' into createsubbuffer 2015-01-16 15:59:24 +01:00
Bastian Köcher
78ce892705 - Added testcase for subbuffer creation 
- Added documentation and asserts for buffer.create_subbuffer function
 2015-01-16 15:57:28 +01:00
Bastian Köcher
1e2da6b9c8 - Added function to create a subbuffer 2015-01-15 15:10:50 +01:00
Kyle Lutz
79aa6758f2 Add wait_guard class 2015-01-07 20:19:05 -08:00
Kyle Lutz
5b5eae8270 Change assert() to static_assert() in copy_async() 
This changes the assert() for contiguous host iterators in
the copy_async() algorithm to a static_assert() so that this
currently unsupported case can be caught at compile-time.
 2014-12-30 20:59:04 -08:00
Kyle Lutz
c74c2abc4f Add four-iterator versions of equal() and mismatch() 2014-12-30 13:44:24 -08:00
Kyle Lutz
d2ea30895d Merge pull request #395 from roshanr95/includes 
Make documentation of includes algorithm more clearer
 2014-12-30 13:40:36 -08:00
Kyle Lutz
a72ee2f289 Merge pull request #394 from roshanr95/bernoulli-distribution 
Change default parameter to 0.5
 2014-12-30 13:39:13 -08:00
Roshan
35b6ff62ee Make documentation of includes algorithm more clearer 2014-12-31 02:00:57 +05:30
Roshan
2a24b282f5 Change default parameter to 0.5 2014-12-31 01:55:55 +05:30
Kyle Lutz
cfdb10c808 Move is_device_iterator to the public API 2014-12-30 10:06:47 -08:00
Kyle Lutz
c31195ec7d Add local_buffer class 2014-12-24 14:06:18 -08:00
Kyle Lutz
417cb03670 Fixes for Mac OS X 2014-12-24 09:25:55 -08:00
Kyle Lutz
e3bbce456b Merge pull request #352 from kylelutz/rename-types-builtin 
Rename types/builtin.hpp to types/fundamental.hpp
 2014-12-21 09:15:29 -08:00
Kyle Lutz
c97c83a562 Rename types/builtin.hpp to types/fundamental.hpp 2014-12-20 13:39:50 -08:00
Kyle Lutz
73b246603b Add for_each_n() algorithm 2014-12-20 13:30:36 -08:00
Kyle Lutz
885a2ad135 Improve documentation 2014-12-20 10:39:38 -08:00
Pavan Yalamanchili
10890fd7d1 BUG FIX: sort_by_key now works with larger vector sizes. 
- sort_by_key now calls radix_sort for size >= 32 even with Compare
 2014-12-15 14:57:49 -05:00
Kyle Lutz
57eec36d20 Release v0.4 2014-12-14 17:03:36 -08:00
Kyle Lutz
c4d49cef20 Merge pull request #341 from kylelutz/device-platform 
Add device::platform() method
 2014-12-14 10:58:38 -08:00
Kyle Lutz
2bab2c5ee7 Add device::platform() method 2014-12-13 11:25:11 -08:00
Kyle Lutz
a4a6614895 Add program_cache to the public API 2014-12-13 10:38:56 -08:00