Grafico dei commit

107 Commit

Autore SHA1 Messaggio Data
Yaxun (Sam) Liu 5c67ee11f4 Fix regression on multi-gpu due to PR#997 2019-04-05 22:54:41 -04:00
Yaxun Sam Liu 8f5c812a68 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.
2019-03-31 10:08:20 -04:00
Wen-Heng (Jack) Chung 04915cea2f 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
2019-03-29 03:45:04 +00:00
Jeff Daily 2845b4c4b8 load program state once per agent 2019-03-27 18:19:10 +00:00
Evgeny 2aa88a4505 tracing callback layer update 2019-03-14 22:43:52 -05:00
Siu Chi Chan fa564a5345 Fix memory leak introduced by previous change to Agent_global.
Make Agent_global manage the lifetime of the name string
2019-03-11 19:51:32 +00:00
Aaron Enye Shi fcde9fe3df Fix Agent_global variables failing hipTestDeviceSymbol
Issue: Header uses std::vector<Agent_global> agent_globals which is created by hip_module.cpp
  - Move iterator fails to copy Agent_global from library source into header version
  - Due to different versions of std::string name in struct Agent_global
Fix: Change Agent_global to use char* name instead of std::string name
2019-03-11 19:51:25 +00:00
Aaron Enye Shi f8aca3159c Fix hash_for undefined reference in hipTestConstant test
Issue: mismatch undefined symbols in different user env
  - Binary expects modified return value std::string&
  - Fails to match libhip_hcc.so: return value is std::string& but doesn't match modified C++ env
Fix: Change return value to char*, create new key std::string in header from char*
2019-03-11 19:51:18 +00:00
Maneesh Gupta 3955f2c131 Merge pull request #949 from gargrahul/single_stream_concurrent_kernels
Add extension for kernel concurrency on same stream
2019-03-06 17:34:54 +05:30
Alex Voicu ed48847237 dlopen() fixes (#929)
* Initial attempt to switch over to internally linked state.

* Add missing CMake update.

* hipLaunchKernelGGLImpl must be inline as well. Ensure internal linkage.

* Ensure global retrieval uses internally linked state.

* Hide HC in the implementation. Minimise ADL woes.

* Strange software exists, and must be catered to.

* Use a less spammy mechanism for ensuring internal linkage / non-export.

* Remove leftover internal detail.
2019-03-06 17:31:44 +05:30
Rahul Garg 263e82a67a Add extension for kernel concurrency on same stream 2019-03-06 12:55:39 +05:30
Wen-Heng (Jack) Chung 6e68d44220 Introduce hash key to HIP module implementation
A hash calculated via FNV-1a algorithm is introduced in ihipModule_t, the
internal of hipModule_t. The hash is used by HIP module APIs such as

- read_agent_global_from_module

to determine whether the agent-scope globals for a module have been iterated.

This commit fixes one issue that applications which load / unload modules
frequently would occasionally fail. After deep investigation of the issue it
turns out the old implementation in read_agent_global_from_module uses
hipModule_t as the key, which is not robust enough, as hipModule_t instances
are allocated dynamically so there are cases that one memory address may be
used by multiple hipModule_t instances. The real solution is to introduce a
uniquely identifiable hash for the code object associated with the HIP module.
And that's the rationale behind this commit.
2019-01-08 17:33:40 +00:00
Yaxun Sam Liu 988dcd1e4a Let hip-clang support --genco 2018-11-27 15:55:50 -05:00
Maneesh Gupta 99bb89b756 Merge pull request #760 from eshcherb/roctracer-hip-frontend-181113
Roctracer hip frontend 181113
2018-11-23 11:08:25 +05:30
Michael Kuron e9b88711e2 Merge branch 'master' into getsymboladdress 2018-11-20 12:03:22 +01:00
Evgeny e362688adf renaming HIP_INIT_CB_API to HIP_INIT_API 2018-11-13 15:33:26 +00:00
Evgeny a13e035fd9 adding lost i the merge change 2018-11-13 15:33:26 +00:00
Evgeny 084a68be63 adding activity prof layer 2018-11-13 15:33:26 +00:00
Rahul Garg 70bb04cc61 Fixed symbol tracking device index 2018-11-13 07:01:17 +05:30
Maneesh Gupta 0bafc38135 Merge pull request #721 from fronteer/my-fix
Make correct checking of the returned hipDeviceptr_t from read_global…
2018-11-08 11:42:08 +05:30
Michael Kuron 4da2d92281 Introduce ihipModuleGetGlobal 2018-11-06 09:54:34 +01:00
Rahul Garg 592efa017f Fixes global symbols tracking in hip_module 2018-10-31 03:22:38 +05:30
Qianfeng Zhang 443698ce00 Make correct checking of the returned hipDeviceptr_t from read_global_description() 2018-10-23 21:13:11 +08:00
Alex Voicu 43fca684c8 Update hip_module.cpp
Typo.
2018-05-18 17:50:45 +01:00
Alex Voicu 40a22d235e Update hip_module.cpp 2018-05-14 17:15:36 +01:00
Alex Voicu eded014abc Don't use magic constants, they're evil.
Also clarify that the register count cannot be queried at the moment.
2018-05-11 11:31:46 +01:00
Alex Voicu bf9529aaa8 Add support for the hipFuncGetAttributes interface. 2018-05-11 03:35:10 +01:00
Rahul Garg c23898f49b hip_module code cleanup
-Fixed missing ihipLogStatus in hipModuleLoad()
-Fixed some ihipXXX functions
2018-04-16 15:35:04 +05:30
Maneesh Gupta 9e47fccc89 Apply .clangformat to all repo source files
Change-Id: I7e79c6058f0303f9a98911e3b7dd2e8596079344
2018-03-12 11:29:03 +05:30
Alex Voicu a704bd8b44 Re-sync with upstream. 2018-02-12 20:20:24 +00:00
Rahul Garg b8c23f979b Fixed host allocated globals address lookup for host usage
Fixed texture driver APIs failure
2018-01-30 18:06:31 +05:30
Rahul Garg d1dcc5025d Fixed build error 2017-12-28 16:15:45 +05:30
Alex Voicu 4e0739c68a This introduces LipoProteinLipase (lpl), a simple tool for creating fat binaries. It represents a direct replacement of the creaky hccgenco.sh script, which had various issues. The format it uses is that of a code object bundle, generated by the Clang Offload Bundler. The output is always suffixed with the ".adipose" extension. It is shared with HCC. The hipcc script and associated tests are modified to use lpl. Help can be obtained by invoking lpl --help. A more computer-sciency / corporate friendly name is likely to be beneficial, which is a reason for choosing easily searchable/replaceable names such as lpl or adipose. 2017-12-08 04:22:57 +00:00
Alex Voicu e186bd9533 This is primarily intended as an additional cleanup of the module functionality, in the aftermath of adopting module based dispatch. The main effort was associated with refactoring the questionable ihipModuleGetSymbol. It was quaintly written and misleading, in that it had little to do with getting symbols, and was exactly retrieving a kernel object. Error handling is modified so as to reduce branching depth. Functions which serve as interfaces to the HSA RT are moved in a separate helper header. Code object readers are properly deleted. Some leftover dead functionality pertaining to associating namespace scope variables with their allocated memory is removed. Executable loading is changed to use a string which holds the ELF image of the code object being loaded, thus avoiding some corner cases where using a istream would fail. 2017-12-03 23:09:06 +00:00
Alex Voicu c6ab6f292b Choose whether or not to use functional grid_launch based on the version of HCC used to compile. 2017-11-29 00:17:44 +00:00
Alex Voicu 6a0efb7ed2 Re-sync with upstream and re-factor platform global management for texture references. 2017-11-28 19:15:29 +00:00
Alex Voicu d37a5a6008 Merge remote-tracking branch 'origin/master' into feature_use_module_based_dispatch_instead_of_pfe
# Conflicts:
#	src/hip_module.cpp
2017-11-28 17:29:11 +00:00
Rahul Garg 38029f2849 Fixed review comments 2017-11-21 21:19:06 +05:30
Alex Voicu f8c1c1b38e Refactor the __device__ versions of memset and memcpy to be less awkward i.e. not return nullptr as opposed to the destination pointer (it can only be assumed it was done for maximum confusion) and actually unroll as they claim to. Change all of the {to, from}Symbol functions to use hipModuleGetGlobal, as opposed to hc::accelerator::get_symbol_address which is no longer valid with module based dispatch. 2017-11-21 02:40:34 +00:00
Rahul Garg eff09fcd86 -Moved coGlobals in hipModule class (takes care of multi module case)
-Used mutex scope for updating coGlobals
2017-11-20 16:23:18 +05:30
Rahul Garg 76f174b536 Update hipModuleGetTexRef API 2017-11-19 22:10:46 +05:30
Alex Voicu 1518611976 Merge remote-tracking branch 'origin/master' into feature_use_module_based_dispatch_instead_of_pfe 2017-11-09 23:43:07 +00:00
Rahul Garg 1851c153f6 Texture driver APIs support 2017-11-09 22:10:55 +05:30
Alex Voicu d12cf0da7d This introduces correct support for agent global variables, and implements hipModuleGetGlobal as an actual equivalent for cuModuleGetGlobal. 2017-11-03 01:44:48 +00:00
Ben Sander 5a80c4108e Merge pull request #245 from scchan/centos_fixes
various fixes for centos/rhel
2017-11-01 18:10:29 +01:00
Alex Voicu 28f87f7d2e This switches HIP from its currently convoluted macro + pfe based dispatch mechanism to a more natural one partially based on the existing module API. The basic idea is that HCC will always correctly emit __global__ functions: as empty-bodied stubs, on host, and as kernels, on device. It then becomes trivial to obtain the mangled name on host, at dispatch, from the function's address, and then to use the mangled name to retrieve the kernel. This should address all problems stemming from serialisation, dubious mismatches due to the manufactured functor, macro-isms et al. It also immediately enables support for generalised globals as a consequence of that being available in the module API. Finally, it will make debug much easier, since the actual names of the __global__ functions will automatically be used in traces etc. One detail is that due to how dispatch works now (hipLaunchKernel and hipLaunchKernelGGL are themselves variadic function templates which deduce the function type of the callee), in certain cases it may be necesssary to insert explicit casts to ensure that the variadic argument list selects a viable overload - this can be observed in some unit tests. Eventually we may be able to remove this limitation, but for now it does not appear terribly onerous. The code is not extremely HIPpie, nor is it fully optimised, but rather is intended as a starting point for the HIP team to make its own. 2017-11-01 15:09:59 +00:00
Siu Chi Chan c1a4ff7b20 Centos/RHEL - remove usage of constexpr since libc++ doesn't enable ctor for constexpr pair in C++11 2017-10-31 18:16:12 +00:00
Ben Sander dc7d993a02 Add ns-level timer for HIP API routines
Refactor some miuses of ihipLogStatus, these should only be in top-level
HIP APIs and should be paired with HIP_API_INIT calls.
2017-10-30 20:20:51 +00:00
Maneesh Gupta 281c543bc3 Make elfio headers private
Change-Id: I3ba174bb46e84a75380207d93a0da6fe3703689e
2017-10-23 10:24:36 +05:30
Alex Voicu b222448efd This fixes incorrect usage of the reader object, which created arcane
mismatches when one reader accessed another's section.
2017-10-09 15:46:38 +01:00