7f21840dc8
added support pinned dma memcpy between host and device
...
[ROCm/clr commit: 620c5c64e6
2016-04-20 14:21:22 -05:00
b857720440
added support for __ldg
...
[ROCm/clr commit: b493eac7e0
2016-04-20 12:25:40 -05:00
57a4505e77
Fixes for P2P and hipDeviceReset
...
- devicereset would lose track of default stream and thus subsequent
synchronization calls might not actually sychronize.
- Also deviceReset now correctly frees streams.
- fix waits in P2P staging copy - first phase (Device0-to-Staging) must
wait for second phase (Staging to Device1) to finish draining the
buffer.
[ROCm/clr commit: 6abfa13c34
2016-04-18 20:49:33 -05:00
66a6cca54d
test update
...
[ROCm/clr commit: 3301deb07c
2016-04-17 07:54:39 -05:00
dd323f5e31
Merge branch 'privatestaging' into p2p
...
Conflicts:
include/hcc_detail/hip_hcc.h
src/hip_hcc.cpp
[ROCm/clr commit: 8d26dfcde3
2016-04-17 06:46:52 -05:00
e71db391ab
added memcpy and p2p test
...
[ROCm/clr commit: e22a910ccc
2016-04-16 17:38:06 -05:00
942509d66e
P2P Update.
...
- add P2P staging buffer copy.
- If copy device does not have sufficient access permissions, fall back
to staging buffer.
- improve docs for which copy device is used.
[ROCm/clr commit: c3bd85595d
2016-04-16 10:18:56 -05:00
419d46f901
Add __ldg
...
[ROCm/clr commit: 0426564a7b
2016-04-16 21:12:09 +08:00
a9401cba2e
improved error returns for hipHostRegister/UnRegister functions
...
[ROCm/clr commit: e8902a6e20
2016-04-15 10:42:31 -05:00
6ee096fbfc
added hipHostRegister/UnRegister api
...
[ROCm/clr commit: 109154f03f
2016-04-15 10:08:10 -05:00
1293709fc0
Replace /opt/hcc -> /opt/rocm/hcc and /opt/hsa -> /opt/rocm/hsa
...
[ROCm/clr commit: cf4c3b0913
2016-04-15 12:56:31 +05:30
76c49bbaf4
Clean up disable.
...
Add USE_HCC_LOCK (disabled)
Disable USE_PEER_TO_PEER.
[ROCm/clr commit: efffb0ed86
2016-04-11 09:09:36 -05:00
f177c8805e
add simple negative P2P tests
...
[ROCm/clr commit: 1acd16eb5d
2016-04-11 07:58:59 -05:00
025938b1de
P2p checkpoint.
...
- set USE_PEER_TO_PEER=3 (requires HCC "am_memtracker_update_peers")
- when enabling peer, turn it on for previously allocated memory.
- hipDeviceCanAccessPeer is no longer self-ware (self does not qualify
as a peer)
- device peerlist always includes self, so when we call allow_access
we never remove self access.
- hipDeviceReset() removes old peer mappings.
[ROCm/clr commit: 9e7efd7c65
2016-04-11 07:58:59 -05:00
ae8ed57c30
Improve P2P test.
...
- Ensure proper device is set before each command.
- Add command line switches:
+bool p_memcpyWithPeer = false; // use the peer device for the P2P copy
+bool p_mirrorPeers = false; // in addition to mapping current to peer
space, map peer to current space.
+int p_peerDevice = -1; // explicly specify which peer to use, else
use p_gpuDevice + 1.
[ROCm/clr commit: b039292a8c
2016-04-11 07:58:58 -05:00
2d90324c09
Improve P2P test.
...
Add option to select which device does the copy.
[ROCm/clr commit: b8e17d52dd
2016-04-11 07:58:58 -05:00
8b827357e9
Fix simple P2P test, had device ptrs swapped
...
[ROCm/clr commit: fcd1d37ff7
2016-04-11 07:58:58 -05:00
a477e65193
Add simple P2P test
...
[ROCm/clr commit: e14b2cd9c3
2016-04-11 07:58:58 -05:00
03eff9ee90
changed c/hip xcompiling source
...
[ROCm/clr commit: 9f7fb0370c
2016-04-11 13:26:09 -05:00
f381fbb2be
added C guard to hip_runtime.h
...
[ROCm/clr commit: 42df534354
2016-04-11 13:01:02 -05:00
302c21d2af
added hipHostRegister support no multi-gpu
...
[ROCm/clr commit: 883954fce2
2016-04-11 10:28:16 -05:00
5a54bbf4d7
added copyrights to experimental/xcompile
...
[ROCm/clr commit: ea4223c812
2016-04-09 14:45:39 -05:00
aad74ffe88
xcompilation tests
...
[ROCm/clr commit: 4f5a8759a5
2016-04-09 14:40:17 -05:00
863bd7d3a4
Directed tests: make HIP_BUILD_LOCAL as default
...
[ROCm/clr commit: 7cbe509b49
2016-04-09 22:05:45 +05:30
b98fcb7764
Added feature for cross-compilation and seperated hcc api from hip api
...
[ROCm/clr commit: 39f27984a8
2016-04-06 14:18:51 -05:00
533c8f6206
Enable directed tests to build HIP locally via cmake option
...
[ROCm/clr commit: fdfedbd60f
2016-04-04 13:53:24 +05:30
c2528bdb57
Revert "Initial (non-default) support for shared library HIP"
...
This reverts commit 8ee3248856a11725cced
2016-04-01 16:07:07 +05:30
8ee3248856
Initial (non-default) support for shared library HIP
...
[ROCm/clr commit: 1c722fa431
2016-04-01 15:45:32 +05:30
82f3f1be67
added device math functions
...
[ROCm/clr commit: c2d04ba59c
2016-03-31 12:29:19 -05:00
522382140d
moved cpp specific device code from headers to source
...
[ROCm/clr commit: 48b8b426d5
2016-03-31 04:45:00 -05:00
7b8ec8fa6d
Added directory for stress testing
...
[ROCm/clr commit: 8d406b68d8
2016-03-31 02:14:12 -05:00
b7a79b9226
Fixed failing tests
...
[ROCm/clr commit: ed17e3489e
2016-03-30 13:00:44 -05:00
266a91d241
C guarding the apis
...
[ROCm/clr commit: c4d93cf3db
2016-03-30 06:08:50 -05:00
7fbdb9fd5c
Merge branch 'privatestaging' of https://github.com/AMDComputeLibraries/HIP-privatestaging into privatestaging
...
[ROCm/clr commit: 00bf37d28d
2016-03-29 11:18:09 -05:00
b8f2f4c45c
rename to hipMultiThreadDevice for consistency
...
[ROCm/clr commit: 3b1cd2d626
2016-03-29 17:33:29 -05:00
875a8a2c84
Test improvements
...
- partition hipThreadSafeDevice into smaller pieces.
- Add debug to hipMultiThreadStream.
- print more precision when mismatch detected.
- enable more tests in CMakeFiles.txt.
[ROCm/clr commit: 98a766ba75
2016-03-29 17:29:31 -05:00
e5ac79daca
Merge branch 'privatestaging' of https://github.com/AMDComputeLibraries/HIP-privatestaging into privatestaging
...
[ROCm/clr commit: de3391637f
2016-03-29 17:12:36 -05:00
d4660e232d
Merge branch 'privatestaging' of https://github.com/AMDComputeLibraries/HIP-privatestaging into privatestaging
...
[ROCm/clr commit: 03c0fcbfec
2016-03-29 14:43:35 -05:00
98f1e31c9b
Calling intrinsics from global kernel
...
[ROCm/clr commit: 3b3db60c25
2016-03-29 11:17:55 -05:00
b22b03fba3
added host flags to hipDeviceBody test
...
[ROCm/clr commit: e652426715
2016-03-29 09:27:57 -05:00
f17131c0d4
Added tests for running device code from a difference source file
...
[ROCm/clr commit: 4320e5c574
2016-03-29 09:22:31 -05:00
d6a2f89b21
Merge branch 'privatestaging' of https://github.com/AMDComputeLibraries/HIP-privatestaging into privatestaging
...
Conflicts:
include/hcc_detail/trace_helper.h
[ROCm/clr commit: ba114a238f
2016-03-29 05:56:03 -05:00
98e2f833b2
Merge branch 'privatestaging' of https://github.com/AMDComputeLibraries/HIP-privatestaging into privatestaging
...
[ROCm/clr commit: 73ecbf5c95
2016-03-29 02:28:20 -05:00
84204d06fe
Not using classes inside headers
...
[ROCm/clr commit: 8e437d0565
2016-03-29 02:27:51 -05:00
4006b37ac5
remove the --hipcc_explicit_lib
...
[ROCm/clr commit: 5e76b13cec
2016-03-28 13:56:23 -05:00
0b0c55af04
Protect _stream_id as well.
...
- move lockedaccessor
- clean up device class.
- add simple ihipDevice constructor.
[ROCm/clr commit: c47b5b04ef
2016-03-26 11:45:25 -05:00
e16e848d55
Make ihipDevice_t thread-safe.
...
Move critical data into separate class and protect with LockAccessor
wrapper class.
For device, the streams list is the critical data since it is modified when
streams are created or destroyed. The streams list is accessed in
several places including when synchronizing across all streams on the
device (ie from the default stream).
Other device data is set once by the device cosntructor and is not critical
so
All functions which acquire the LockAccessor now named with "locked_" prefix.
[ROCm/clr commit: 4dd77c6612
2016-03-26 10:46:20 -05:00
c5f87744bb
Added test for C + HIP
...
[ROCm/clr commit: 026aaf987f
2016-03-25 13:45:28 -05:00
4a292b8c70
updated hipdevicesync test
...
[ROCm/clr commit: d6bc773d5b
2016-03-25 06:41:49 -05:00
ba8c181e21
added functional tests for hip device apis
...
[ROCm/clr commit: 65510d6e4c
2016-03-25 05:49:33 -05:00
Aditya Atluri
Ben Sander
streamhsa
Maneesh Gupta
scchan