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.
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.
This adds a program cache which can be used by algorithms and other
functions to store programs which may be re-used. This improves
performance by reducing the need for costly recompilation of commonly
used programs.
Program caches are context specific and multiple copies of the same
context will use the same program cache. They are created and accessed
by the global get_program_cache() function.
For now, only a few algorithms and functions (radix sort, mersenne
twister, fixed size sorts) make use of the program cache.
This adds a new macro for the unit-tests which checks a range of
values on the device against an array of values on the host. This
simplifies writing tests and removes the need to explicitly copy
values back to the host for verification.
refs kylelutz/compute#9
device, context, and queue are initialized statically in `context_setup.hpp`.
With this change all tests are able to complete when an NVIDIA GPU is in
exclusive compute mode.
Side effect of the change:
Time for all tests to complete reduced from 15.71 to 13.03 sec Tesla C2075.
