AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
6,604 Commits 3 Branches 0 Tags
537b3dd67faed63a2e6ba494cf440fffcea2b09b
Commit Graph

6604 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
David Salinas d1f4e7ea54 append the ELF flags for sram-ecc and xnack to the target triple per code object 
[ROCm/clr commit: 4d0dc45078]
 2019-04-05 13:17:11 -04:00
Evgeny Mankov 9eb10c5b4d [HIPIFY][doc] Fix typos and minor text inaccuracies in HIP FAQ 
[ROCm/clr commit: 4990a1003b]
 2019-04-05 19:19:38 +03:00
foreman a0f0714729 P4 to Git Change 1766264 by cpaquot@cpaquot-ocl-lc-lnx on 2019/04/05 11:38:18 
SWDEV-145570 - [HIP] Use a context with all devices in system for host register
	hipHostRegister and hipMemcpy 0x10 and 0x20 fail in mGPU systems because
	we only register the memory on the current device. But in HIP, the registering
	needs to happen on all devices.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_context.cpp#17 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_internal.hpp#26 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#50 edit


[ROCm/clr commit: 7a58581c81]
 2019-04-05 11:58:25 -04:00
foreman c70ae33d9e P4 to Git Change 1766115 by cpaquot@cpaquot-ocl-lc-lnx on 2019/04/04 18:29:12 
SWDEV-145570 - Don't assert when LOG_LEVEL=3 and return value isn't HIP_SUCCESS.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_internal.hpp#25 edit


[ROCm/clr commit: 54f50c23fc]
 2019-04-04 18:41:09 -04:00
foreman fa9f13db43 P4 to Git Change 1766106 by kjayapra@99_HIPWS_SLV_CHECKIN on 2019/04/04 18:07:26 
SWDEV-144570 - Implementation of hipMemcpyToSymbol and simillar fns for PAL.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_internal.hpp#24 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#22 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#26 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/devprogram.hpp#23 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palprogram.cpp#89 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palprogram.hpp#37 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocprogram.cpp#102 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocprogram.hpp#46 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/memory.cpp#135 edit


[ROCm/clr commit: d0c8e12d00]
 2019-04-04 18:22:40 -04:00
foreman df8687b499 P4 to Git Change 1766076 by michliao@hliao-dev-00-hip.rocm-workspace on 2019/04/04 17:35:28 
SWDEV-144570 - Fix pointer attribute query.

	- Device should be numbered from zero.

	RB: http://ocltc.amd.com/reviews/r/17090/

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#49 edit


[ROCm/clr commit: 3b461e01ae]
 2019-04-04 17:50:08 -04:00
Siu Chi Chan f6837a4e7f minor workaround for lazy binding 
[ROCm/clr commit: b5045af7e9]
 2019-04-02 17:28:06 -04:00
Jenkins 8b22d2b1f6 Merge 'master' into 'amd-master' 
Change-Id: I4369829444ebd150e59f07f5db44e4584d9cc396


[ROCm/clr commit: 5a6d260117]
 2019-04-02 12:33:16 -05:00
Anthony Cowley 191dfc8dff hipify: translate cudaErrorInvalidSymbol 
[ROCm/clr commit: f15dd58a91]
 2019-04-02 08:54:37 -04:00
Maneesh Gupta bedff2406a Merge pull request #998 from yxsamliu/doc 
hip-clang: update installation guide.

[ROCm/clr commit: 2d9e5615b7]
 2019-04-02 05:08:18 +00:00
Maneesh Gupta 9d2c0c8e5d Merge pull request #997 from yxsamliu/mgpu 
hip-clang: fix kernel not found on multi-gpu

[ROCm/clr commit: a6e7dfefe3]
 2019-04-02 05:07:31 +00:00
Evgeny Mankov 1168e2925b Merge pull request #1001 from emankov/master 
[HIPIFY][tests] Fix typo in test for CUDA 10.x

[ROCm/clr commit: 887c096ed7]
 2019-04-01 19:53:42 +03:00
Evgeny Mankov 17ac0d3e19 [HIPIFY][tests] Fix typo in test for CUDA 10.x 
[ROCm/clr commit: 5b59f87305]
 2019-04-01 19:52:08 +03:00
Evgeny Mankov 7d93455770 Merge pull request #1000 from emankov/master 
[HIPIFY][cmake] Update CMakeLists and Readme cause CUDA 10.1 and clang 8.0.0 are released

[ROCm/clr commit: 16128da4a3]
 2019-04-01 19:49:46 +03:00
Evgeny Mankov 236624a229 [HIPIFY][cmake] Update CMakeLists and Readme cause CUDA 10.1 and clang 8.0.0 are released 
[ROCm/clr commit: 799a6f5512]
 2019-04-01 19:44:52 +03:00
foreman 1251ab0181 P4 to Git Change 1764069 by michliao@hliao-dev-00-hip-workspace on 2019/04/01 11:23:34 
SWDEV-144570 - Handle zero-byte memset & memcpy

	- Properly handle zero-byte memset & memcpy by skipping the real stuff.

	RB: http://ocltc.amd.com/reviews/r/17062/

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#48 edit


[ROCm/clr commit: 6220c5059d]
 2019-04-01 11:51:06 -04:00
Yaxun Sam Liu 12ac74bad1 hip-clang: fix kernel not found on multi-gpu 
__hipRegisterFunction is called during by .init functions during program initialization.
It calls hipModuleGetFunction to locate kernel symbol in code objects. hipModuleGetFunction
assumes current device when locating kernel symbols. This works for HCC but not for hip-clang,
since hip-clang needs to locate kernel symbols for different devices without switching
between devices.

This patch introduces a new hsa agent parameter to ihipModuleGetFunction, which allows
__hipRegisterFunction to choose the correct hsa agent when locating kernel symbols. By
default it uses this_agent(), therefore this patch has no impact on HCC.


[ROCm/clr commit: 8f5c812a68]
 2019-03-31 10:08:20 -04:00
Yaxun (Sam) Liu 54c2b79351 Update INSTALL.md 
[ROCm/clr commit: 76a9fdd924]
 2019-03-30 08:29:08 -04:00
Yaxun Sam Liu 79ac0097dc hip-clang: update installation guide. 
[ROCm/clr commit: 598604aa7f]
 2019-03-30 08:24:49 -04:00
Jenkins d02a41f291 Merge 'master' into 'amd-master' 
Change-Id: Id0b95bac00a76909d68f53406875028fdbb97ac7


[ROCm/clr commit: 45c92008c1]
 2019-03-29 04:09:52 -05:00
Wen-Heng (Jack) Chung cfe930f9d6 Make hipModuleGetGlobal be in HIP runtime so it can be discovered at runtime (#981) 
* Make hipModuleGetGlobal be in HIP runtime so it can be discovered at runtime

In HIP PR #929, quite a few HIP public APIs were made as inline functions with
hidden visibility. It was necessary to support applications with shared
libraries with GPU kernels launched via hipLaunchKernelGGL(), after HIP runtime
is initialized.

In empirical tests, the implementation has been proved to be a bit too
excessive, especially for hipModuleGetGlobal(). The function is used by another
type of client applications which relies on the existence of this function
within HIP runtime so global symbols from HSA code objects loaded dynamically
at runtime can be retrieved programmtically.

This commit moves hipModuleGetGlobal() back to src/hip_module.cpp, and makes it
visible and not inline, to fulfill requirements for applications
aforementioned. It does not change the behavior of applications depending on
hipLaunchKernelGGL().

* Add HIP_INIT_API into the implementation of hipModuleGetGlobal

Address review comments.

* Fix failing HIP unit tests


[ROCm/clr commit: 04915cea2f]
 2019-03-29 03:45:04 +00:00
Maneesh Gupta d99bc4c540 Merge pull request #992 from gargrahul/handle_d2d_memcpy2d 
Handle D2D in memcpy2D

[ROCm/clr commit: f9f4cee347]
 2019-03-28 04:41:36 +00:00
Maneesh Gupta a2191a80a6 Merge pull request #991 from jeffdaily/per_agent_program_state 
load program state once per agent

[ROCm/clr commit: 79f24a1553]
 2019-03-28 04:40:46 +00:00
Jeff Daily 5233d41c6c improve program state commentary 
Disambiguate calling many varibles "agent".
More detail in exception message.
Create and discard map placeholders; no need to call std::vector::clear() on map value.


[ROCm/clr commit: f5e4fff6cc]
 2019-03-27 21:40:27 +00:00
Rahul Garg 73bb9a74bb Handle D2D in memcpy2D 
[ROCm/clr commit: 50d623981e]
 2019-03-28 02:21:45 +05:30
Jeff Daily 9cee2c5311 load program state once per agent 
[ROCm/clr commit: 2845b4c4b8]
 2019-03-27 18:19:10 +00:00
foreman 7b86800c59 P4 to Git Change 1762349 by yaxunl@yaxunl-lc10 on 2019/03/27 11:50:12 
SWDEV-145570 - Workaround for mismatch of device name and bundle id for gfx906.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#25 edit


[ROCm/clr commit: dbb7c04ac1]
 2019-03-27 12:53:17 -04:00
Jenkins 6f87eff3aa Merge 'master' into 'amd-master' 
Change-Id: Id7c377bd22abccbdffd9d1f8bd31a482102813ce


[ROCm/clr commit: 4d25945c7d]
 2019-03-27 04:09:48 -05:00
Maneesh Gupta a0b29c8ed0 Merge pull request #987 from gargrahul/fix_hostmalloc_double_device_map 
Avoid double mapping of devices to hostMalloc buffer

[ROCm/clr commit: 93906a072c]
 2019-03-27 05:23:47 +00:00
Maneesh Gupta 6fb7f626ba Merge pull request #990 from mhbliao/hliao/master/sw 
SWDEV-184380 Fix hcc compilation

[ROCm/clr commit: 178e3ecdca]
 2019-03-27 05:23:26 +00:00
Maneesh Gupta 80543778a3 Merge pull request #986 from mhbliao/hliao/master/build 
[hip] Fix typo in macro `hipLaunchKernel`

[ROCm/clr commit: 3bcb3a9239]
 2019-03-27 05:22:44 +00:00
Michael LIAO c5717a37d7 SWDEV-184380 Fix hcc compilation 
- `hcc` has no builtin. Need to invoke LLVM intrinsic directly.


[ROCm/clr commit: d355122bf9]
 2019-03-26 15:20:17 -04:00
Rahul Garg 0d47ae4203 Let hipHostMalloc always share/map pinned host ptr 
[ROCm/clr commit: 9b38380c03]
 2019-03-26 10:19:13 +05:30
Rahul Garg 21d7bbab11 Avoid double mapping of devices to hostMalloc buffer 
[ROCm/clr commit: ad11972f47]
 2019-03-25 23:07:05 +05:30
Michael LIAO 5482fa8102 [hip] Fix typo in macro hipLaunchKernel 
[ROCm/clr commit: 13655df76e]
 2019-03-25 12:06:46 -04:00
Jenkins ab30e62eb9 Merge 'master' into 'amd-master' 
Change-Id: I759c530530011ee9cff5aef00944c190e4204df2


[ROCm/clr commit: 17cfee36d7]
 2019-03-25 09:17:20 -05:00
Jenkins ca2958f7e3 Merge 'master' into 'amd-master' 
Change-Id: Ie456ccae78b50086939e1eaf65a5f768d40b4aca


[ROCm/clr commit: 956084082d]
 2019-03-25 04:09:54 -05:00
Maneesh Gupta 817e064745 Merge pull request #970 from mangupta/swdev-172995 
hipExtMallocWithFlags implementation

[ROCm/clr commit: c20d233585]
 2019-03-25 07:46:53 +00:00
Maneesh Gupta 9e2774e81e Merge pull request #962 from gargrahul/add_2d_copy_fallback 
Add 2D fallback to use copy kernel

[ROCm/clr commit: 9de28dfa5a]
 2019-03-25 07:46:43 +00:00
Rahul Garg 66ce9921d5 2D Fallback needs hcc workweek 19101 or higher 
[ROCm/clr commit: bec3995700]
 2019-03-25 12:07:28 +05:30
Maneesh Gupta 505fc1e98c hipExtMallocWithFlags needs hcc workweek 19115 or higher 
[ROCm/clr commit: 45255ab492]
 2019-03-25 11:41:20 +05:30
Maneesh Gupta 888b43cc6f Merge pull request #982 from ROCm-Developer-Tools/hack_swdev-173477 
HACK for SWDEV-173477

[ROCm/clr commit: 158eac9374]
 2019-03-22 09:14:38 +00:00
Maneesh Gupta 4224de0edc Merge pull request #980 from ntrost57/master 
fixed loss of accuracy in hipCfma()

[ROCm/clr commit: f3966c6cb1]
 2019-03-22 09:14:09 +00:00
Jenkins 35adb52d38 Merge 'master' into 'amd-master' 
Change-Id: Id44416aee5dc592d7b6351926abf53d4a49e9b45


[ROCm/clr commit: 03789ce057]
 2019-03-22 04:09:52 -05:00
Wen-Heng (Jack) Chung 86379d694f HACK for SWDEV-173477 
For code objects with global symbols of length 0, ROCR runtime would
ignore them even though they exist in the symbol table. Therefore the
result from read_agent_globals() can't be trusted entirely.

As a workaround to tame applications which depend on the existence of
global symbols with length 0, always return hipSuccess here.

This behavior shall be reverted once ROCR runtime has been fixed to
address SWDEV-173477


[ROCm/clr commit: cf7ad0f184]
 2019-03-21 17:18:16 +00:00
foreman adb9f1ffe3 P4 to Git Change 1759609 by kjayapra@1_HIPWS_SL_IPC on 2019/03/21 12:09:27 
SWDEV-144570 - Disbaling __hipHostRegister implementation to unblock pytorch tests in PAL.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#24 edit


[ROCm/clr commit: 95d1962401]
 2019-03-21 12:43:25 -04:00
Nico Trost 0b3f8dce2b fixed loss of accuracy in hipCfma() 
[ROCm/clr commit: 725486fb11]
 2019-03-21 10:30:10 +01:00
eshcherb 0cf8b184a5 adding hip_prof_gen verbose log (#977) 
* adding hip_prof_gen verbose log

* adding stderr fatal error

* adding no error exit by default

* adding hip_prof_str regeneration dependencies

* adding more informative messages

* fixing error mesage


[ROCm/clr commit: 045c6afa2c]
 2019-03-21 05:28:18 +00:00
foreman 707f3ec6ed P4 to Git Change 1759248 by cpaquot@cpaquot-ocl-lc-lnx on 2019/03/20 18:32:19 
SWDEV-183452 - [HIP] Typo of | instead of &

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#47 edit


[ROCm/clr commit: 9682f81a75]
 2019-03-20 19:18:10 -04:00
foreman 95c6f8fed5 P4 to Git Change 1759224 by cpaquot@cpaquot-ocl-lc-lnx on 2019/03/20 17:47:43 
SWDEV-183452 - [HIP] PyTorch test_pin_memory fails
	hipPointerGetAttributes should return hipMemoryTypeHost for hipHostMalloc memories.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#46 edit


[ROCm/clr commit: a63b38c304]
 2019-03-20 18:43:07 -04:00