Ben Sander 806e58ca63 Refactor asyncCopy and syncCopy to fix deadlock case. ...

- Minimize time that locks are held.
- Eliminate copy code that locked stream and ctx at same time.
    - Stream was locked to ensure thread-safe enqueue to the queue.
    - Devices were locked to query peer-lists.

Change-Id: Ibe8880bb7fb995a3da8f90ff911f212d81525018

2016-09-27 15:45:40 -05:00

..

cuda

Initial commit for GPUOpen Launch

2016-01-26 20:14:33 -06:00

hcc_acc.h

Updates docs for hipHcc* functions, move to header

2016-09-22 13:05:47 -05:00

hip_blas.h

Adding hipblas include files

2016-08-12 11:59:25 +05:30

hip_fp16.h

moved half support to a source file

2016-06-30 18:23:29 +05:30

hip_hcc.h

Refactor asyncCopy and syncCopy to fix deadlock case.

2016-09-27 15:45:40 -05:00

hip_ldg.h

fix_ldg

2016-06-09 16:56:05 -04:00

hip_runtime_api.h

Doxygen documentation changes for hipCtxXXX APIs

2016-09-26 23:11:07 +05:30

hip_runtime.h

Fix fence functions doxygen group

2016-09-16 15:06:54 +05:30

hip_texture.h

Doxygen related document updates for texture/ array related functions

2016-09-26 16:18:53 +05:30

hip_util.h

added CR and hipsetdeviceflags func

2016-03-24 07:04:01 -05:00

hip_vector_types.h

Add compiler annotations to make_ functions

2016-06-16 16:18:35 +05:30

hipComplex.h

Added nvcc path for hipComplex APIs

2016-08-22 10:29:46 -05:00

host_defines.h

Grid-launch updates to 2.0 and cleanup of old.

2016-06-21 23:24:38 -05:00

trace_helper.h

improve debug for event

2016-09-02 09:46:59 -05:00