Граф коммитов

174 Коммитов

Автор SHA1 Сообщение Дата
Sun, Peng 2ca9eec9fc Add more include header file for GGL, to make it self-compilable
Change-Id: I833cb194784450fb86e7961a7f9fe196ce3c7da5
2017-04-03 11:16:35 -05:00
Maneesh Gupta cea13c34c3 Enable promote free HCC by default
Change-Id: I5f82a8d958dd675a8a46d9d17458c71321daab7c
2017-04-03 11:18:30 +05:30
Sun, Peng a7d7797781 Fix warpSize, for related issue in hipeigen and torch
Change-Id: Ic66b24923a363304dca189011869ba7a0a6f8895
2017-04-01 15:42:53 -05:00
Sun, Peng df4cee1a57 GGL update, fix for thread-safe access to streams (accelerator_views).
Change-Id: I6dd329a85b3ba7de23d52823febee0c53857a981
2017-04-01 14:50:39 -05:00
Aditya Atluri 3859f59189 Fixed copyright and header names
Change-Id: Id595c65ea3b7289e87be4c42db5d8a31905a4fdd
2017-03-31 12:40:29 -05:00
Aditya Atluri 614e7db513 fixed header names
Change-Id: I21650d6398187d3767b28e8ac81b2642d3b89a0e
2017-03-31 12:18:55 -05:00
Aditya Atluri 1cead6a4cd added new api hipHccModuleLaunchKernel
1. hipHccModuleLaunchKernel is same as hipModuleLaunchKernel with OpenCL workitem model
2. Added copy right
3. Fixed header naming

Change-Id: I6a7c35a3566e2f8d3f5056613e34193775d4b236
2017-03-31 12:11:34 -05:00
Sun, Peng e3a4547703 remove extra GGL header info
Change-Id: I09f0d1b64a7a31eb2e926f19b69b7bafbacc7787
2017-03-30 18:16:56 -05:00
Sun, Peng 9983fdb9e2 Rename hipLaunchKernelV3 to hipLaunchKernelGGL
Change-Id: I303daae006db41e9b04eb591e0b09b2717a7cf66
2017-03-30 18:10:17 -05:00
Sun, Peng c081099de1 Enable GGL as the default kernel launch method
Change-Id: I8022d126ee28ff7e4d9a96209e399d4243d39d8b
2017-03-30 13:36:46 -05:00
Sun, Peng a32f50b5c7 Update GGL to fix one Torch build issue
Change-Id: I95a2a335902e3c368ed29f075ac72eabbb64c97e
2017-03-29 09:03:21 -05:00
sunway513 6f38d0f55e resolve GGL merge conflicts
Change-Id: I7a5ec3696cf2dd1a77b1686536a1cb84cbfed66e
2017-03-27 22:46:41 -05:00
Rahul Garg 41f0ebebf0 Fix for MemcpyFromSymbol on HIP/NVCC path
Change-Id: Ice38307f72870ae468cbf0861e104f0fa46dfd56
2017-03-27 00:35:10 +05:30
Rahul Garg 33a8cdc6d3 Added support for Primary Context Management APIs
Change-Id: I70f91b4492e112dd8e12ecf511fdc18a27944a06
2017-03-26 23:45:54 +05:30
Sun, Peng 4ca09e977d update GGL implementation to use hipLaunchKernel
Change-Id: Ibc08185c814bb07d54f3e68016b10eb7b9f2bf4b
2017-03-21 15:46:56 -05:00
Sun, Peng 6b9be30b70 Update GGL implementation to extended overload set for make_lambda_wrapper
Change-Id: I949f113671ddf155db8689e8a7f23d415839a7b5
2017-03-21 15:46:44 -05:00
Sun, Peng 14f8c7cf74 revert workaround for square sample and update doc on GGL
Change-Id: I731c68ca4111e7dc2e45bef51c4cad2c23fc81f8
2017-03-21 10:26:09 -05:00
Sun, Peng e21d4f03f1 merge Alex' GGL fix for non-specialized kernel function launch
Change-Id: Idbf7ca669c38ee5c0f654bcabdd1b498abb29f69
2017-03-20 16:34:24 -05:00
Aditya Atluri 52859a8a40 added support for lgammaf and lgamma
1. Implementation inside HIP

Change-Id: I657263b7276a57c56081d3336fef816b5f204eff
2017-03-17 18:26:10 -05:00
Sun, Peng 2585812fba Disable additional debug warning message
Change-Id: Ic5c374589bfad387a7c4c5346430a490e2c6e2a7
2017-03-17 15:03:03 -05:00
pensun faf0fbebe9 Change the #define of GENERIC_GRID_LAUNCH to take valueat compilation, disable warning messages
Change-Id: Ic6c011529e26de359bcda1e7083727e7ee52887b
2017-03-17 14:59:34 -05:00
pensun ad882222b0 Initial integration with Alex' Generic Grid Launch
Change-Id: I559afb80e9e39ec0d119bb3bf3b85ef9e448caf6
2017-03-17 14:59:34 -05:00
Ben Sander 20ea5e179f Move USE_PROMOTE_FREE_HCC 2017-03-17 12:04:13 -05:00
Ben Sander 92fd4c8f10 Add __device__ to needful functions for promote-free. 2017-03-17 11:19:48 -05:00
Ben Sander f4256cb57c Add USE_PROMOTE_FREE_HCC for smooth transition to new HCC caps.
ADDRESS_SPACE_1 defines
2017-03-17 11:04:39 -05:00
Ben Sander c9f64bbd2d Add simple device-side assert macro
Currently swallows asserts but will compile.
2017-03-17 10:25:56 -05:00
Maneesh Gupta e3aa3d7ba6 Merge branch 'amd-master' into amd-develop
Change-Id: I01a77c34e11b1b02cb20e11b105741f6bd383543
2017-03-15 09:56:45 +05:30
Rahul Garg 09d1a46139 hipMemsetD8 support for HIP/NVCC path
Change-Id: I48eee8266afd7b45a12d5ce2c4849b687a006c0f
2017-03-14 23:49:21 +05:30
Rahul Garg dccf9e2aa9 Added hipMemsetD8
Change-Id: I6a230a036c9c46c72a77d5f93c16ce8a00c3f837
2017-03-14 22:11:34 +05:30
Wen-Heng (Jack) Chung 17af837120 Revert "Changes to HIP to cope with Promote-free HCC"
This reverts commit 0005dd5f66.

Change-Id: I20a9bab3883ad09913b320210344d37599cb8fcd
2017-03-14 22:59:27 +08:00
Maneesh Gupta 0853c8ac2e Merge branch 'amd-master' into amd-develop 2017-03-14 13:44:41 +05:30
Aditya Atluri af56898ea0 make sure the inter-thread intrinsics are working post hawaii
Change-Id: I30ea5284c2160276f5bc0f937dfd386ca8640ce8
2017-03-13 11:16:05 -05:00
Ben Sander b7acb85fa8 Refactor registered memory calls. 2017-03-11 09:18:27 -06:00
Aditya Atluri a47066153f fixed warning raised by g++ using hip_vector_types.h
Change-Id: I9e7cdfc8b28b03b690eecd068529cf7629296d68
2017-03-10 15:14:26 -06:00
Rahul Garg 3af487007b IPC supported using ROCR APIs
Change-Id: I0a353b1240098f4b20fa266a871f5f5826290af9
2017-03-10 23:45:28 +05:30
pensun 1a2844e3a3 update porting guide for updated __HIP_DEVICE_COMPILE__
Change-Id: I0f025d354f76e2d728231bf112a77e8c8fcacc8c
2017-03-10 10:01:12 -06:00
Aditya Atluri 046ec0375b Added architecture guards around __shfl, dpp and ds_permute device functions
Change-Id: I10f9b08618fbf25b61c1932278fc5759e41c0d66
2017-03-10 08:40:59 -06:00
pensun 14a5d3c80d add inline to all hip_complex operators
Change-Id: Ifba5966c297cbc9299c39ecfc45c7296003ebb5d
2017-03-08 14:06:13 -06:00
Aditya Atluri 7b7d53f875 Added new API, hipMemPtrGetInfo
1. This API returns memory allocation size of pointer
2. Added test to check its functionality

Change-Id: I87976d817b5a6ca5530336c09e7cb0420601cb2c
2017-03-07 13:46:29 -06:00
Wen-Heng (Jack) Chung 0005dd5f66 Changes to HIP to cope with Promote-free HCC
Squashed commit of the following:

commit c111b5bd10d7c2a5b0b1ad8b07f6e81185b47b39
Author: Wen-Heng (Jack) Chung <whchung@gmail.com>
Date:   Sat Mar 4 17:06:46 2017 +0800

    Use __device__ for all variables and functions to be used in kernel path

    Abolish __device and adopt [[hc]] in HIP implementation, so __device__ can be
    used on all HIP applications, no matter they are variables or functions.

    Change-Id: I20ca25857ce3bc3e42a5ebf65cafea2c8492f4c7

commit 30c0e4e4701bbf6bd9a7182e0320a71ff73d3a83
Author: Wen-Heng (Jack) Chung <whchung@gmail.com>
Date:   Thu Mar 2 12:14:11 2017 +0800

    XXX FIXME get around LDS spills caused in Promote-free HCC

    hipDynamicShared2 uses all 64KB of LDS for computation. But in Promote-free HCC
    there are cases where LDS spills would occur, which would make the test case to
    hang.

    In this workaround commit we reduce the size of dynamic LDS used to get around
    this known issue, and will revert this commit when LDS spills are resolved in
    HCC.

    Change-Id: If648b36200a4f9143951a8129192bcb7ed0bef5e

commit e803173be2d73e2f132a7ff7f61e7a20b4083d34
Author: Wen-Heng (Jack) Chung <whchung@gmail.com>
Date:   Wed Mar 1 21:41:41 2017 +0800

    Fix math functions which take pointer arguments

    Change-Id: I332c997e640edbc44824691e2a9434c6b3dadefa

commit de590c469e213c42090ff83dbd060f25bb1d6047
Author: Wen-Heng (Jack) Chung <whchung@gmail.com>
Date:   Wed Mar 1 18:38:54 2017 +0800

    Changes to cope with Promote-free HCC

    - abolish usage of address_space GNU attribute
    - use __device in file-scope global variables which would be accessed by GPU kernels
    - temporarily disable some math functions which take pointer arguments

    Change-Id: I730311dee848e20e763e35cd3980317fce0dce0d

Change-Id: I1f6b970b53b9401eeaaab08f04a7b9fed0fb8cf0
2017-03-08 01:32:59 +08:00
Aditya Atluri 9f575721aa added new field to hipDeviceProp_t structure gcnArch.
1. It is an integer containing gfx values 701, 801, 802, 803
2. On NV path, it is zero

Change-Id: I2b4c7f48981d0214d8c6b1905d2cc85b16203419
2017-03-07 11:24:32 -06:00
Ben Sander 8e50134d31 Modify memcpy and memset to follow C/C++ standard:
- memcpy src is const
    - memset val is int
2017-03-06 16:38:22 -06:00
Evgeny Mankov 430856ca35 [HIP] [FIX] Memcpy Async functions on nvcc path.
+ in hipMemcpyDtoDAsync: cuMemcpyDtoD -> cuMemcpyDtoDAsync
+ in hipMemcpyDtoHAsync: cuMemcpyDtoH -> cuMemcpyDtoHAsync

P.S.
"The types CUstream and cudaStream_t are identical and may be used interchangeably", thus explicit c-like type cast is not needed, aka CUstream(stream).
2017-03-01 23:04:34 +03:00
Aditya Atluri 37bc29e44b changed __half enabling from 4 to >3
Change-Id: Id974c6d5326e87a4c5941f831c5bb2747cdebd2d
2017-02-28 17:13:29 -06:00
pensun be95cc914f Define __HIPCC__ flag at compile time when using HIPCC on HCC path
Change-Id: I5e967e0e2327264d5d3b0ca705c2504fcd33d75e
2017-02-28 16:20:48 -06:00
Evgeny Mankov 47a27e6b17 [HIP] Add missing Device attribute on nvcc path.
+ missing cudaDevAttrComputeCapabilityMinor case as added for hipDeviceGetAttribute query for hipDeviceAttributeComputeCapabilityMinor.
2017-02-28 18:40:13 +03:00
pensun c5ae8c423a remove extra spaces for hip_common.h platform defines
Change-Id: Ie0e39256abba307429985371671cde01f5ea2cc9
2017-02-27 12:55:22 -06:00
Maneesh Gupta 528736cf58 Fix nvcc path samples that include math_functions.h
Change-Id: I94bb577b93983535178d8f0dcae57aaa72871534
2017-02-27 13:20:05 +05:30
Aditya Atluri 7ac5017cb9 Added initial support for hipMemcpyFromSymbol. But not working!
Change-Id: I48d8c7de4ec9f85c6c942be995fb488a3931f5d7
2017-02-23 11:29:06 -06:00
Aditya Atluri 3d348b2d81 added runtime api hipMemcpyFromSymbolAsync
Change-Id: Ibaf925faf0ba464dd0ed6c5ea74c224c2ce38889
2017-02-22 19:16:35 -06:00