Update: Add complete suppport for OpenCL 1.2 atomics on arrays by map…

[CoreRasurae]

…ping Java Arrays of AtomicIntegers (refs #81)

Provides support for full OpenCL 1.2 atomics on array of integers as described in #81

[Issue81AtomicsSupport 8a0692f] Update: Add complete suppport for OpenCL 1.2 atomics on arrays by mapping Java Arrays of AtomicIntegers (refs #81) 8 files changed, 1803 insertions(+), 10 deletions(-) create mode 100644 src/test/java/com/aparapi/runtime/Issue81AtomicsSupportTest.java

[codecov-io]

Codecov Report

Merging #82 into master will decrease coverage by 0.43%. The diff coverage is 11.59%.

@@             Coverage Diff              @@
##             master      #82      +/-   ##
============================================
- Coverage     46.59%   46.15%   -0.44%     
- Complexity      849      852       +3     
============================================
  Files            57       57              
  Lines          9464     9593     +129     
  Branches       1539     1568      +29     
============================================
+ Hits           4410     4428      +18     
- Misses         4606     4706     +100     
- Partials        448      459      +11

[freemo]

@CoreRasurae I presume this should work with opencl 2.0 and 2.1 atomics as well?

[CoreRasurae]

Yes, this should work with opencl 2.0 and 2.1 capable devices, however the atomics specification for OpenCL 2.0 and 2.1 is more complete and this implementation does not cover that. I am working on a few improvements for code quality... should I open a new PR for that? Also I noticed that the feature #79 has a memory leak in the unit tests, as I'm not disposing the kernels. Should I do a PR for both?

[freemo]

@CoreRasurae yes please open those as PR, I will get them into a 1.6.1 release ASAP.

On Mon, Apr 16, 2018 at 9:10 AM, CoreRasurae notifications@github.com wrote:

Yes, this should work with opencl 2.0 and 2.1 capable devices, however the atomics specification for OpenCL 2.0 and 2.1 is more complete and this implementation does not cover that. I am working on a few improvements for code quality... should I open a new PR for that? Also I noticed that the feature #79 https://github.com/Syncleus/aparapi/issues/79 has a memory leak in the unit tests, as I'm not disposing the kernels. Should I do a PR for both?

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/Syncleus/aparapi/pull/82#issuecomment-381593639, or mute the thread https://github.com/notifications/unsubscribe-auth/AC5JAkC11owQ7MIXzCDFUtgueL5GDh1iks5tpJhXgaJpZM4TVrn5 .