Sarunya Pumma a68ea730c2 Remove device mapping from shareWithAll memory ...

When shareWithAll memory (e.g., host memory) is allocated, set appId
in hc::AmPointerInfo to -1 to indicate that this memory is not mapped
to any device.  Peer checking in ihipStream_t::canSeeMemory is not
necessary if memory is shared with all devices.  Thus, it is skipped.

Note that earlier host memory is always mapped to device 0 and HIP
always performs peer checking for all kinds of hipMemcpy.  Since the
peer checking process requires context locking, hipMemcpy from/to host
memory always grabs device 0's context lock.  Therefore, if there is
another thread holding the context lock of device 0 (e.g.,
hipDeviceSynchronize on device 0), hipMemcpy will have to wait for the
lock until it can actually perform memcpy.  This can significantly
deteriorate execution performance.

Signed-off-by: Sarunya Pumma <sarunya.pumma@amd.com>


[ROCm/hip commit: 8111fd3b8b]

2018-07-28 23:15:16 -07:00

..

elfio

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

AMDGPUPTNote.h

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

AMDGPURuntimeMetadata.h

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

code_object_bundle.cpp

Fix build failure in code_object_bundle.cpp

2018-07-10 16:49:59 -04:00

device_util.cpp

Merge branch 'master' into move-memcpy

2018-07-17 10:51:42 +05:30

device_util.h

Implement Memory Fence Functions in header

2018-07-04 23:35:24 +00:00

env.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

env.h

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

functional_grid_launch.inl

It is unclear what I was thinking when authoring the original code...

2018-07-17 14:04:57 +01:00

grid_launch.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

hip_clang.cpp

Change HIP fat binary magic number

2018-05-17 17:04:51 -04:00

hip_context.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

hip_db.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

hip_device.cpp

Add integrated device property

2018-06-02 13:11:16 +05:30

hip_error.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

hip_event.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

hip_hcc_internal.h

Update hip_hcc_internal.h

2018-07-04 09:33:51 +05:30

hip_hcc.cpp

Remove device mapping from shareWithAll memory

2018-07-28 23:15:16 -07:00

hip_memory.cpp

Remove device mapping from shareWithAll memory

2018-07-28 23:15:16 -07:00

hip_module.cpp

Update hip_module.cpp

2018-05-18 17:50:45 +01:00

hip_peer.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

hip_stream.cpp

callback handler: don't need to wait for the thread to become ready

2018-06-02 17:55:37 -04:00

hip_surface.cpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

hip_surface.h

Added hip_surface header file

2018-03-01 12:22:56 +05:30

hip_texture.cpp

Null check before setting offset

2018-07-24 12:25:40 +05:30

hip_texture.h

HIP Texture Support

2017-07-17 15:16:12 -04:00

hip_util.h

Fixed copyright and header names

2017-03-31 12:40:29 -05:00

hsa_helpers.hpp

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30

macro_based_grid_launch.inl

[cmake] Added target static_check for static code analysis

2017-12-14 12:16:44 +05:30

program_state.cpp

Keep the map which tracks GPU kernel symbols to grow monotonically

2018-06-18 16:54:18 -05:00

trace_helper.h

Apply .clangformat to all repo source files

2018-03-12 11:29:03 +05:30