AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
5,142 Commits 3 Branches 0 Tags
8a42ac4d03c7a88f7c16da51dce15dee080f0ca3
Commit Graph

4437 Commits

Author SHA1 Message Date
foreman cf855653b9 P4 to Git Change 1778949 by michliao@hliao-dev-00-hip.rocm-workspace on 2019/05/06 16:36:53 
SWDEV-144570 - Fix build failure after switching to gcc-7

	- Hex representation of float needs gnu++11. We'd better not relying on
	  that. Change the float in hex format into alternative representation.

	RBT: http://ocltc.amd.com/reviews/r/17300/

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#29 edit


[ROCm/hip commit: 09920e6024]
 2019-05-06 17:20:00 -04:00
Maneesh Gupta f7eaff474b Implement hipExtGetLinkTypeAndHopCount for ROCm devices 
Change-Id: Ie5bb4f640ac6d189c7fceeab22627a7494fd10bd


[ROCm/hip commit: 035ef04e1f]
 2019-05-06 15:54:31 +05:30
Michael LIAO d71327882c [test] Add device variant of std::declval. 
- Current clang disallows any invocation of wrong-side functions even
  under context with type-inspection only. Work around that by adding a
  variant of `std::decl` with `__device__` attribute.


[ROCm/hip commit: 4fbfb646f6]
 2019-05-03 15:58:31 -04:00
Michael LIAO cb55dc384a [test] Use explicit cast for address space cast. 
[ROCm/hip commit: 2746d95906]
 2019-05-03 14:56:00 -04:00
Maneesh Gupta f49a4c6d89 Merge pull request #1062 from mhbliao/hliao/master/icmp 
[hip] Re-implement ballot using AMDGCN builtins

[ROCm/hip commit: 117bdd8774]
 2019-05-03 17:48:19 +05:30
Maneesh Gupta f42dc08384 Merge pull request #1058 from mhbliao/hliao/master/devfunc 
[Device Function] Fix implementation

[ROCm/hip commit: 37d01a7da9]
 2019-05-03 17:47:51 +05:30
Konstantin Pyzhov ae391b3211 hipMultiThreadDevice test: Reduced maximum number of created HIP streams on Windows. 
[ROCm/hip commit: e29c9d50a6]
 2019-05-03 05:43:30 -04:00
Konstantin Pyzhov 45eb04ea5c Refined hipSetDevice test. 
[ROCm/hip commit: 069fb7e486]
 2019-05-03 05:39:45 -04:00
Michael LIAO b7ae0b6442 [test] Remove unused common routines. 
- Remove used common routines, which variadic functions, not supported
  feature, are used.


[ROCm/hip commit: 16a27e06ee]
 2019-05-02 12:34:07 -04:00
Michael LIAO 2c2b897902 [devfunc] Add necessary __device__ and __host__ attributes. 
- Minor clean up to keep consistent function declaration.


[ROCm/hip commit: 44692978fe]
 2019-05-01 22:26:35 -04:00
foreman 8c0263f6c5 P4 to Git Change 1777225 by cpaquot@cpaquot-ocl-lc-lnx on 2019/05/01 18:47:25 
SWDEV-188233 - [HIP] Implemented hipGetChannelDesc.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_texture.cpp#11 edit


[ROCm/hip commit: 3ad9c3f984]
 2019-05-01 19:53:29 -04:00
foreman 486506f113 P4 to Git Change 1777158 by cpaquot@cpaquot-ocl-lc-lnx on 2019/05/01 17:12:01 
SWDEV-188219 - [HIP] Add hipStreamCreateWithPriority. NOP right now.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_device_runtime.cpp#14 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.def.in#13 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_hcc.map.in#14 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_stream.cpp#17 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/commandqueue.hpp#21 edit


[ROCm/hip commit: ba93e1ae21]
 2019-05-01 18:43:47 -04:00
emankov 1f57ee6125 [HIPIFY][tests] Add cuSPARSE CSR-BCSR-SPMV-conversions example 
[ROCm/hip commit: e3082f5142]
 2019-04-30 17:37:34 +03:00
Michael LIAO e637e72364 [Device Function] Fix implementation of __bitinsert_u64 
- It's a common mistake by assuming 1 << shamt would be promoted to
  64-bit, if shamt is a 64-bit integer. That's not the case. Replace
  that left shift to a 64-bit one to ensure it won't fall into undefined
  behavior.
- Fix the host-side implementation as well for device function testing.


[ROCm/hip commit: 9bd2d5746d]
 2019-04-30 08:59:13 -04:00
Michael LIAO cc4de2bc28 [devfunc] Re-implement ballot using AMDGCN builtins 
- As the signature of `amdgcn.icmp` is changed for next-gen chip, using
  clang builtins is portable way to hide that details.


[ROCm/hip commit: a64637da2c]
 2019-04-29 17:21:25 -04:00
Michael LIAO 448be14118 [hipcc] Repace --rpath with --rpath-link 
- With `--rpath`, app's RUNPATH is set to the temporary library path
  during build. Instead, `--rpath-link` should be used to leave RUNPATH
  untouched. User should set LD_LIBRARY_PATH or system ld.so.conf to
  proper setting.


[ROCm/hip commit: 7f8b80103d]
 2019-04-29 16:01:16 -04:00
foreman 1363c10d20 P4 to Git Change 1775995 by gandryey@gera-w8 on 2019/04/29 13:46:53 
SWDEV-79445 - OCL generic changes and code clean-up
	- Enable P2P extension for PAL path, currently it's staging copy only
	- Fix P2P staging copy

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#55 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#243 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.hpp#336 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/paldevice.cpp#130 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palsettings.cpp#77 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palvirtual.cpp#131 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palvirtual.hpp#59 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocdevice.cpp#124 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocdevice.hpp#36 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocvirtual.cpp#74 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/command.cpp#92 edit


[ROCm/hip commit: 345f3f989a]
 2019-04-29 13:58:26 -04:00
Evgeny Mankov 2b16e01c66 [HIPIFY][doc] Update Readme.md: latest cuDNN 7.5.1.10 is supported 
+ tested with CUDA 9.0, 9,2, 10.0 and 10.1


[ROCm/hip commit: c0705f892b]
 2019-04-29 15:41:08 +03:00
Aaron Enye Shi 62e96cb4cf Revert "Use COMgr to read Kernel Args Metadata (#1006)" 
This reverts commit 882006555b.


[ROCm/hip commit: a3d118eaa8]
 2019-04-26 16:04:56 -04:00
Aaron Enye Shi c797e6e27b Revert "Add COMGR relative path for build machines" 
This reverts commit 01a11c94f8.


[ROCm/hip commit: 48701ad4ba]
 2019-04-26 16:04:56 -04:00
Aaron Enye Shi 3585159140 Revert "Add dependency on amd_comgr in hip-config-*.cmake.in" 
This reverts commit b9526364bb.


[ROCm/hip commit: 59a5965fe1]
 2019-04-26 16:04:56 -04:00
foreman cb5f5dd0bb P4 to Git Change 1775366 by kjayapra@99_HIPWS_SLV_CHECKIN on 2019/04/26 14:48:23 
SWDEV-144570 - hipFuncGetAttributes() API changes.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_internal.hpp#29 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_module.cpp#23 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#28 edit


[ROCm/hip commit: 9011aa9282]
 2019-04-26 15:15:48 -04:00
Maneesh Gupta b9526364bb Add dependency on amd_comgr in hip-config-*.cmake.in 
Change-Id: Iac1d851a8cfb99224e9c5926780273d9b9b08426


[ROCm/hip commit: ef99ffd9f4]
 2019-04-25 15:26:33 -04:00
Sameer Sahasrabuddhe c3db7b929a minor cleanup: eliminate repetition 
[ROCm/hip commit: abb9375707]
 2019-04-25 20:41:16 +05:30
Evgeny Mankov d8f4eff1ef [HIPIFY][perl][fix][258] Memory fence device functions are supported now 
[ROCm/hip commit: abd1c53cf8]
 2019-04-25 13:27:30 +03:00
Evgeny Mankov 86d68901da [HIPIFY][DNN] cudnnSetFilter4dDescriptor support 
[ROCm/hip commit: 3fee0f3765]
 2019-04-25 12:18:51 +03:00
Evgeny Mankov d51679ed8e [HIPIFY][fix][#204] Suppress warning message: #pragma once in main file 
[ROCm/hip commit: 6d3c443234]
 2019-04-24 20:35:52 +03:00
Evgeny Mankov 43839c780e [HIPIFY][doc] Update README.md 
+ A few words about clang patches to work with CUDA 9.2 - 10.0 on Windows;
+ Fix cuDNN versions with correct values.


[ROCm/hip commit: 4651dce3f0]
 2019-04-24 17:40:35 +03:00
Maneesh Gupta c36aa3a056 Merge pull request #1043 from mhbliao/hliao/master/fp16 
[hip] Fix including of hip_fp16.h

[ROCm/hip commit: ffe9f86fe8]
 2019-04-24 16:50:46 +05:30
Maneesh Gupta 7938012378 Merge pull request #1042 from mhbliao/hliao/master/ldg 
[hip] Fix use of `__HIP_CLANG_ONLY__` in `hip_ldg.h`.

[ROCm/hip commit: de6c680767]
 2019-04-24 16:50:37 +05:30
Maneesh Gupta 7b5ed45968 Merge pull request #1040 from eshcherb/roctracer-hip-frontend-190422 
hip_prof_api.h include under __cplusplus

[ROCm/hip commit: e489f7579a]
 2019-04-24 16:50:27 +05:30
Maneesh Gupta 5235da375e Merge pull request #1039 from gargrahul/fix_ptrgetattr_nvcc 
Fix hipPointerGetAttributes for NVCC

[ROCm/hip commit: 2975221560]
 2019-04-24 16:50:18 +05:30
Rahul Garg c01236f679 Add hipMallocManaged default functional support (#1036) 
* Add hipMallocManaged default functional support

* Fix build error

* Add dtest


[ROCm/hip commit: 2bc2c46d4d]
 2019-04-24 16:50:03 +05:30
Maneesh Gupta 4d9977c10b Merge pull request #1034 from kpyzhov/master 
Minor fixes for 64-bit device functions.

[ROCm/hip commit: a016777acb]
 2019-04-24 16:49:36 +05:30
Maneesh Gupta 72c7388422 Merge pull request #1031 from yxsamliu/fix-init 
Fix missing arg in HIP_INIT_API

[ROCm/hip commit: 8dc8c58ddd]
 2019-04-24 16:49:23 +05:30
Maneesh Gupta b17642bb10 Merge pull request #1028 from gargrahul/fix_d2d_async_test 
[dtest] Fix D2DAsync test

[ROCm/hip commit: eed5928007]
 2019-04-24 16:49:13 +05:30
Aaron Enye Shi 01a11c94f8 Add COMGR relative path for build machines 
[ROCm/hip commit: 920fe246d7]
 2019-04-23 17:16:26 -04:00
Evgeny Mankov 3fb77dbef5 [HIPIFY][doc] Provide patches for clang's bug 38811 
+ Update Readme.md accordingly


[ROCm/hip commit: 57931b3056]
 2019-04-23 21:13:00 +03:00
Evgeny Mankov a2b6b1e7e9 [HIPIFY][hipify-perl] Formatting 
[ROCm/hip commit: defc6f8155]
 2019-04-23 17:55:47 +03:00
Michael LIAO 6eec9d4ac8 [hip] Fix including of hip_fp16.h 
- Separate the definition of `__HCC_OR_HIP_CLANG__`, `__HCC_ONLY__`, and
  `__HIP_CLANG_ONLY__` into hip_common.h so that it could be included in
  hip_fp16.h, which may be included separately in app.


[ROCm/hip commit: dc0d7bd5ce]
 2019-04-23 09:16:00 -04:00
Michael LIAO 27aed8c172 [hip] Fix use of __HIP_CLANG_ONLY__ in hip_ldg.h. 
- Check its value instead of whether it's defined or not.


[ROCm/hip commit: 6fb07acc8c]
 2019-04-22 23:22:32 -04:00
Evgeny 17474d2b58 hip_prof_api.h include under __cplusplus 
[ROCm/hip commit: af3f3ccb2b]
 2019-04-22 21:14:18 -05:00
Rahul Garg be49602b29 Fix hipPointerGetAttributes for NVCC 
[ROCm/hip commit: 69a3d6b72a]
 2019-04-23 03:22:25 +05:30
foreman 17f613bfde P4 to Git Change 1772785 by cpaquot@cpaquot-ocl-lc-lnx on 2019/04/22 14:31:18 
SWDEV-144570 - [HIP] Lazy build kernels to avoid overfilling dev memory.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_internal.hpp#28 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#27 edit


[ROCm/hip commit: 80fb280643]
 2019-04-22 15:19:24 -04:00
foreman b0a31b621c P4 to Git Change 1772349 by mshivama@mshivama_tf on 2019/04/20 07:35:24 
SWDEV-187020 - Basic "Hello World" Tensorflow program fails to execute on HIP/VDI/PAL/LINUX

	Make sure that the size of `dest` is big enough to hoid `src` including trailing zero byte

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_device.cpp#19 edit


[ROCm/hip commit: 84679ca20e]
 2019-04-21 21:50:04 -04:00
foreman 1e8f8da0d9 P4 to Git Change 1772193 by mshivama@mshivama_tf on 2019/04/19 09:39:39 
SWDEV-187020 -  Basic "Hello World" Tensorflow program fails to execute on HIP/VDI/PAL/LINUX

	the API hiDeviceGetName() should pass the length of device name as argument to strncpy() instead
	of total (dest) memory size which is being passed to it by its caller. Also, bit of a code clean-up.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_device.cpp#18 edit


[ROCm/hip commit: 7a1d02c8a1]
 2019-04-19 09:49:37 -04:00
Konstantin Pyzhov 2e0c67bbd3 Fix for __popcll() device function implementation. 
[ROCm/hip commit: beadaab661]
 2019-04-19 08:53:22 -04:00
Yaxun (Sam) Liu d4bce6c019 Fix missing arg in HIP_INIT_API 
[ROCm/hip commit: bb5c620b13]
 2019-04-18 16:18:31 -04:00
Konstantin Pyzhov 53b232fb2a Fix for __ffsll() device functions. 
[ROCm/hip commit: b7bd29924a]
 2019-04-18 13:07:24 -04:00
foreman ac0387e22f P4 to Git Change 1771752 by cpaquot@cpaquot-ocl-lc-lnx on 2019/04/18 11:54:18 
SWDEV-145570 - [HIP] Use isAsync flag for async cases.

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#54 edit


[ROCm/hip commit: 17b5ecc6dc]
 2019-04-18 12:15:25 -04:00