In case when the tile size is greater than the number of active threads,
the coalesced group size should be equal to the number of active threads.
Change-Id: I1d41322f2428a07862a590cb5d34b01243383b7c
The warpSize variable is set to the value of the __AMDGCN_WAVEFRONT_SIZE macro,
which is a meaningless default in host code.
The resolution for SWDEV-449015 will introduce diagnostics for uses of this
macro in host code, which includes the current definition of the warpSize
variable. With the __device__ specifier, the definition of the warpSize
variable will not cause these diagnostics.
This change does not stop the variable from being used in host code since clang
intentionally does not diagnose uses of __device__ constexpr variables in host
code.
Change-Id: I0317217affe94fdf2dfd9ad0f134e68f5173245f
AMD CPUs have had avx512_bf16 support for quite some time now (from
consumer Ryzen 7000 series to enterprise grade CPUs). This
patch should allow users to use the hardware bf16 unit when running the
__host__ variants of the function. This can be enabled via `hipcc ...
-mavx512vl -mavx512bf16`.
Change-Id: I67c377afc95ddfe8d45a048dce078a247d4a1878
- enforce incrementing the table versioning number when a table size changes outside of ifdef for ROCPROFILER_REGISTER
- add new HIP_ENFORCE_ABI entries
- update the HipDispatchTable size and bump HIP_RUNTIME_API_TABLE_STEP_VERSION to 1
- re-enable rocprofiler-register
Change-Id: Ie0cc1d8491c5640056e5dd393ea243e4dce4e8a9
- Add API table versioning defines to hip_api_trace.hpp
- Add rocprofiler-register implementation in hip_api_trace.cpp
- created ToolInit implementation
- moved some local functions into anonymous namespace
- renamed some function to be overloads
- made most of the initialization code templated (reduced code duplication)
- enforce ABI
- Updated hipamd/src/CMakeLists.txt
- find rocprofiler-register package (enabled by default but not required)
- Address review comments
- simplify size calculation for dispatch table
- remove setting .size field in GetDispatchTableImpl
- Fix windows build
- Do not enforce ABI on Windows
Change-Id: I08766e8a083528a1236996274bf4522e0e8a1b9f
The following builtins from the CUDA spec are implemented:
- __all_sync, __any_sync, __ballot_sync and __activemask
- __match_any_sync and __match_all_sync
- __shfl_sync, __shfl_up_sync, __shfl_down_sync, and __shfl_xor_sync
The following builtins are NOT implemented, pending support in the compiler:
- __reduce_add_sync, __reduce_min_sync, __reduce_max_sync
- __reduce_and_sync, __reduce_or_sync, __reduce_xor_sync
Change-Id: I07dedbbfe5449f4b5c9b040bed59f5603ccec8c3
It cannot be moved to amd_device_functions.h because that causes circular
dependences when trying to use the macro in other files. So we create a new
header and move all assert/abort macros to that common header.
As a side-effect, also fix the macro to correctly expand the entire condition
argument, and also consume the trailing semicolon.
Change-Id: I43688c8e61183503a3a1a039b91321a3779152af
If the compiler decides not to inline these functions, we might break ODR (one definition rule) due to this file being included in multiple files and being linked together
Change-Id: Iacbfdabb53f5b4e5db8c690b23f3730ec9af16c0
This reverts commit 353dbe6e3b.
Reason for revert: This is considered a breaking change and requires
multiple apps to change their behavior. This will be reintroduced in later releases.
Change-Id: I3481627115af1872785585a155cc6a0ecfbe1372
Marko Arandjelovic
sdashmiz
Jatin Chaudhary
Ioannis Assiouras
Fabian Ritter
Saleel Kudchadker
Anusha GodavarthySurya
Sameer Sahasrabuddhe
kjayapra-amd
Satyanvesh Dittakavi
Rahul Manocha
Jonathan R. Madsen
Jaydeep Patel
jiabaxie
jujiang
Rahul Garg