Skip to content

Commit

Permalink
merge amd-staging into amd-mainline
Browse files Browse the repository at this point in the history 
Change-Id: I535bd80160437e5a9e399d2aafab4e6e388e2abb
  • Loading branch information
@emankov
emankov committed Oct 13, 2024
2 parents f63b359 + e2258cd commit f156b77
Show file tree
Hide file tree
Showing 35 changed files with 2,174 additions and 1,040 deletions.
22 changes: 21 additions & 1 deletion CHANGELOG.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,27 @@
Documentation for HIPIFY is available at
[https://rocmdocs.amd.com/projects/HIPIFY/en/latest/](https://rocmdocs.amd.com/projects/HIPIFY/en/latest/).

## HIPIFY for ROCm 6.2.2
## HIPIFY for ROCm 6.3.0

### Additions

* CUDA 12.6.1 support
* cuDNN 9.5.0 support
* LLVM 19.1.1 support
* `rocBLAS` 64-bit APIs support
* Initial support for direct hipification of `cuDNN` into `MIOpen` under the `--roc` option
* Initial support for direct hipification of `cuRAND` into `rocRAND` under the `--roc` option
* [#1650] Added a filtering ability for the supplementary hipification scripts

### Fixes

* Correct `roc` header files support

### Known issues

* [#1617] Support for `fp8` data types

## HIPIFY for ROCm 6.2.4

### Additions

Expand Down
373 changes: 233 additions & 140 deletions bin/hipify-perl
View file

Large diffs are not rendered by default.

113 changes: 57 additions & 56 deletions docs/hipify-clang.rst
View file
Original file line number Diff line number Diff line change
Expand Up @@ -37,7 +37,7 @@ Dependencies

* `LLVM+Clang <http://releases.llvm.org>`_ of at least version
`4.0.0 <http://releases.llvm.org/download.html#4.0.0>`_; the latest stable and recommended release:
`19.1.0 <https://github.com/llvm/llvm-project/releases/tag/llvmorg-19.1.0>`_.
`19.1.1 <https://github.com/llvm/llvm-project/releases/tag/llvmorg-19.1.1>`_.

* `CUDA <https://developer.nvidia.com/cuda-downloads>`_ of at least version
`7.0 <https://developer.nvidia.com/cuda-toolkit-70>`_, the latest supported version is
Expand Down Expand Up @@ -184,7 +184,8 @@ Dependencies
- `12.3.2 <https://developer.nvidia.com/cuda-12-3-2-download-archive>`_
- ✅
- ✅
* - `19.1.0 <https://github.com/llvm/llvm-project/releases/tag/llvmorg-19.1.0>`_:sup:`4`
* - `19.1.0 <https://github.com/llvm/llvm-project/releases/tag/llvmorg-19.1.0>`_,
`19.1.1 <https://github.com/llvm/llvm-project/releases/tag/llvmorg-19.1.1>`_:sup:`4`
- `12.6.1 <https://developer.nvidia.com/cuda-downloads>`_:sup:`4`
- **Latest stable config**
- **Latest stable config**
Expand Down Expand Up @@ -227,7 +228,7 @@ Dependencies
In most cases, you can get a suitable version of ``LLVM+Clang`` with your package manager. However, you can also
`download a release archive <http://releases.llvm.org/>`_ and build or install it. In case of multiple versions of ``LLVM`` installed, set
`CMAKE_PREFIX_PATH <https://cmake.org/cmake/help/latest/variable/CMAKE_PREFIX_PATH.html>`_ so that
``CMake`` can find the desired version of ``LLVM``. For example, ``-DCMAKE_PREFIX_PATH=D:\LLVM\19.1.0\dist``.
``CMake`` can find the desired version of ``LLVM``. For example, ``-DCMAKE_PREFIX_PATH=D:\LLVM\19.1.1\dist``.

Usage
============================================================
Expand Down Expand Up @@ -260,7 +261,7 @@ header files used during the hipification process:

.. code:: shell
./hipify-clang square.cu --cuda-path=/usr/local/cuda-12.6 --clang-resource-directory=/usr/llvm/19.1.0/dist/lib/clang/19
./hipify-clang square.cu --cuda-path=/usr/local/cuda-12.6 --clang-resource-directory=/usr/llvm/19.1.1/dist/lib/clang/19
For more information, refer to the `Clang manual for compiling CUDA <https://llvm.org/docs/CompileCudaWithLLVM.html#compiling-cuda-code>`_.

Expand Down Expand Up @@ -397,7 +398,7 @@ To ensure LLVM being found or in case of multiple LLVM instances, specify the pa

.. code-block:: bash
-DCMAKE_PREFIX_PATH=/usr/llvm/19.1.0/dist
-DCMAKE_PREFIX_PATH=/usr/llvm/19.1.1/dist
On Windows, specify the following option for CMake in the first place:
``-G "Visual Studio 17 2022"``.
Expand Down Expand Up @@ -471,7 +472,7 @@ LLVM <= 9.0.1
LLVM >= 10.0.0
-----------------

1. Download `LLVM project <https://github.com/llvm/llvm-project/releases/tag/llvmorg-19.1.0>`_ sources.
1. Download `LLVM project <https://github.com/llvm/llvm-project/releases/tag/llvmorg-19.1.1>`_ sources.

2. Build `LLVM project <http://llvm.org/docs/CMake.html>`_:

Expand Down Expand Up @@ -545,7 +546,7 @@ LLVM >= 10.0.0

.. code-block:: shell
-DCUDA_DNN_ROOT_DIR=D:/CUDA/cuDNN/9.4.0
-DCUDA_DNN_ROOT_DIR=D:/CUDA/cuDNN/9.5.0
5. [Optional] Install `CUB 1.9.8 <https://github.com/NVIDIA/cub/releases/tag/1.9.8>`_ for ``CUDA < 11.0`` only;
for ``CUDA >= 11.0``, the CUB shipped with CUDA will be used for testing.
Expand Down Expand Up @@ -574,13 +575,13 @@ LLVM >= 10.0.0

.. code-block:: bash
python /usr/llvm/19.1.0/llvm-project/llvm/utils/lit/setup.py install
python /usr/llvm/19.1.1/llvm-project/llvm/utils/lit/setup.py install
**Windows**:

.. code-block:: shell
python D:/LLVM/19.1.0/llvm-project/llvm/utils/lit/setup.py install
python D:/LLVM/19.1.1/llvm-project/llvm/utils/lit/setup.py install
In case of errors similar to ``ModuleNotFoundError: No module named 'setuptools'``, upgrade the ``setuptools`` package:

Expand All @@ -594,23 +595,23 @@ LLVM >= 10.0.0

.. code-block:: bash
-DLLVM_EXTERNAL_LIT=/usr/llvm/19.1.0/build/bin/llvm-lit
-DLLVM_EXTERNAL_LIT=/usr/llvm/19.1.1/build/bin/llvm-lit
**Windows**:

.. code-block:: shell
-DLLVM_EXTERNAL_LIT=D:/LLVM/19.1.0/build/Release/bin/llvm-lit.py
-DLLVM_EXTERNAL_LIT=D:/LLVM/19.1.1/build/Release/bin/llvm-lit.py
* ``FileCheck``:

**Linux**:

Copy from ``/usr/llvm/19.1.0/build/bin/`` to ``CMAKE_INSTALL_PREFIX/dist/bin``.
Copy from ``/usr/llvm/19.1.1/build/bin/`` to ``CMAKE_INSTALL_PREFIX/dist/bin``.

**Windows**:

Copy from ``D:/LLVM/19.1.0/build/Release/bin`` to ``CMAKE_INSTALL_PREFIX/dist/bin``.
Copy from ``D:/LLVM/19.1.1/build/Release/bin`` to ``CMAKE_INSTALL_PREFIX/dist/bin``.

Alternatively, specify the path to ``FileCheck`` in the ``CMAKE_INSTALL_PREFIX`` option.

Expand All @@ -637,16 +638,16 @@ On Linux, the following configurations are tested:

* Ubuntu 14: LLVM 4.0.0 - 7.1.0, CUDA 7.0 - 9.0, cuDNN 5.0.5 - 7.6.5
* Ubuntu 16-19: LLVM 8.0.0 - 14.0.6, CUDA 7.0 - 10.2, cuDNN 5.1.10 - 8.0.5
* Ubuntu 20-21: LLVM 9.0.0 - 19.1.0, CUDA 7.0 - 12.6.1, cuDNN 5.1.10 - 9.4.0
* Ubuntu 22-23: LLVM 13.0.0 - 19.1.0, CUDA 7.0 - 12.6.1, cuDNN 8.0.5 - 9.4.0
* Ubuntu 20-21: LLVM 9.0.0 - 19.1.1, CUDA 7.0 - 12.6.1, cuDNN 5.1.10 - 9.5.0
* Ubuntu 22-23: LLVM 13.0.0 - 19.1.1, CUDA 7.0 - 12.6.1, cuDNN 8.0.5 - 9.5.0

Minimum build system requirements for the above configurations:

* CMake 3.16.8, GNU C/C++ 9.2, Python 3.0.

Recommended build system requirements:

* CMake 3.30.2, GNU C/C++ 13.2, Python 3.12.5.
* CMake 3.30.4, GNU C/C++ 13.2, Python 3.12.7.

Here's how to build ``hipify-clang`` with testing support on ``Ubuntu 23.10.01``:

Expand All @@ -656,10 +657,10 @@ Here's how to build ``hipify-clang`` with testing support on ``Ubuntu 23.10.01``
-DHIPIFY_CLANG_TESTS=ON \
-DCMAKE_BUILD_TYPE=Release \
-DCMAKE_INSTALL_PREFIX=../dist \
-DCMAKE_PREFIX_PATH=/usr/llvm/19.1.0/dist \
-DCMAKE_PREFIX_PATH=/usr/llvm/19.1.1/dist \
-DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-12.6.1 \
-DCUDA_DNN_ROOT_DIR=/usr/local/cudnn-9.4.0 \
-DLLVM_EXTERNAL_LIT=/usr/llvm/19.1.0/build/bin/llvm-lit \
-DCUDA_DNN_ROOT_DIR=/usr/local/cudnn-9.5.0 \
-DLLVM_EXTERNAL_LIT=/usr/llvm/19.1.1/build/bin/llvm-lit \
../hipify
The corresponding successful output is:
Expand All @@ -683,20 +684,20 @@ The corresponding successful output is:
-- - Test hipify-clang : ON
-- - Is part of HIP SDK : OFF
-- Found ZLIB: /usr/lib/x86_64-linux-gnu/libz.so (found version "1.2.13")
-- Found LLVM 19.1.0:
-- - CMake module path : /usr/llvm/19.1.0/dist/lib/cmake/llvm
-- - Clang include path : /usr/llvm/19.1.0/dist/include
-- - LLVM Include path : /usr/llvm/19.1.0/dist/include
-- - Binary path : /usr/llvm/19.1.0/dist/bin
-- Found LLVM 19.1.1:
-- - CMake module path : /usr/llvm/19.1.1/dist/lib/cmake/llvm
-- - Clang include path : /usr/llvm/19.1.1/dist/include
-- - LLVM Include path : /usr/llvm/19.1.1/dist/include
-- - Binary path : /usr/llvm/19.1.1/dist/bin
-- Linker detection: GNU ld
-- ---- The below configuring for hipify-clang testing only ----
-- Found Python: /usr/bin/python3.12 (found version "3.12.5") found components: Interpreter
-- Found Python: /usr/bin/python3.12 (found version "3.12.7") found components: Interpreter
-- Found lit: /usr/local/bin/lit
-- Found FileCheck: /GIT/LLVM/trunk/dist/FileCheck
-- Initial CUDA to configure:
-- - CUDA Toolkit path : /usr/local/cuda-12.6.1
-- - CUDA Samples path :
-- - cuDNN path : /usr/local/cudnn-9.4.0
-- - cuDNN path : /usr/local/cudnn-9.5.0
-- - CUB path :
-- Found CUDAToolkit: /usr/local/cuda-12.6.1/targets/x86_64-linux/include (found version "12.6.68")
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
Expand All @@ -705,7 +706,7 @@ The corresponding successful output is:
-- Found CUDA config:
-- - CUDA Toolkit path : /usr/local/cuda-12.6.1
-- - CUDA Samples path : OFF
-- - cuDNN path : /usr/local/cudnn-9.4.0
-- - cuDNN path : /usr/local/cudnn-9.5.0
-- - CUB path : /usr/local/cuda-12.6.1/include/cub
-- Configuring done (0.5s)
-- Generating done (0.0s)
Expand All @@ -722,11 +723,11 @@ The corresponding successful output is:
Running HIPify regression tests
===============================================================
CUDA 12.6.68 - will be used for testing
LLVM 19.1.0 - will be used for testing
LLVM 19.1.1 - will be used for testing
x86_64 - Platform architecture
Linux 6.5.0-15-generic - Platform OS
64 - hipify-clang binary bitness
64 - python 3.12.5 binary bitness
64 - python 3.12.7 binary bitness
===============================================================
-- Testing: 106 tests, 12 threads --
Testing Time: 6.91s
Expand Down Expand Up @@ -818,16 +819,16 @@ Tested configurations:
- ``3.11.4``
* - ``17.0.1`` :sup:`6` - ``18.1.8`` :sup:`7`
- ``7.0 - 12.3.2``
- ``8.0.5 - 9.4.0``
- ``2019.16.11.38, 2022.17.10.5``
- ``3.30.0``
- ``3.12.5``
* - ``19.1.0``
- ``8.0.5 - 9.5.0``
- ``2019.16.11.40, 2022.17.11.4``
- ``3.30.4``
- ``3.12.7``
* - ``19.1.1``
- ``7.0 - 12.6.1``
- ``8.0.5 - 9.4.0``
- ``2019.16.11.38, 2022.17.10.5``
- ``3.30.0``
- ``3.12.5``
- ``8.0.5 - 9.5.0``
- ``2019.16.11.40, 2022.17.11.4``
- ``3.30.4``
- ``3.12.7``

:sup:`5` LLVM 14.x.x is the latest major release supporting Visual Studio 2017.

Expand All @@ -852,55 +853,55 @@ Building with testing support using ``Visual Studio 17 2022`` on ``Windows 11``:
-DHIPIFY_CLANG_TESTS=ON \
-DCMAKE_BUILD_TYPE=Release \
-DCMAKE_INSTALL_PREFIX=../dist \
-DCMAKE_PREFIX_PATH=D:/LLVM/19.1.0/dist \
-DCMAKE_PREFIX_PATH=D:/LLVM/19.1.1/dist \
-DCUDA_TOOLKIT_ROOT_DIR="C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.6" \
-DCUDA_SDK_ROOT_DIR="C:/ProgramData/NVIDIA Corporation/CUDA Samples/v12.5" \
-DCUDA_DNN_ROOT_DIR=D:/CUDA/cuDNN/9.4.0 \
-DLLVM_EXTERNAL_LIT=D:/LLVM/19.1.0/build/Release/bin/llvm-lit.py \
-DCUDA_DNN_ROOT_DIR=D:/CUDA/cuDNN/9.5.0 \
-DLLVM_EXTERNAL_LIT=D:/LLVM/19.1.1/build/Release/bin/llvm-lit.py \
../hipify
The corresponding successful output is:

.. code-block:: shell
-- Selecting Windows SDK version 10.0.22621.0 to target Windows 10.0.22631.
-- The C compiler identification is MSVC 19.39.33523.0
-- The CXX compiler identification is MSVC 19.39.33523.0
-- The C compiler identification is MSVC 19.41.34120.0
-- The CXX compiler identification is MSVC 19.41.34120.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: C:/Program Files/Microsoft Visual Studio/2022/Community/VC/Tools/MSVC/14.39.33519/bin/Hostx64/x64/cl.exe - skipped
-- Check for working C compiler: C:/Program Files/Microsoft Visual Studio/2022/Community/VC/Tools/MSVC/14.41.34120/bin/Hostx64/x64/cl.exe - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: C:/Program Files/Microsoft Visual Studio/2022/Community/VC/Tools/MSVC/14.39.33519/bin/Hostx64/x64/cl.exe - skipped
-- Check for working CXX compiler: C:/Program Files/Microsoft Visual Studio/2022/Community/VC/Tools/MSVC/14.41.34120/bin/Hostx64/x64/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- HIPIFY config:
-- - Build hipify-clang : ON
-- - Test hipify-clang : ON
-- - Is part of HIP SDK : OFF
-- Found LLVM 19.1.0:
-- - CMake module path : D:/LLVM/19.1.0/dist/lib/cmake/llvm
-- - Clang include path : D:/LLVM/19.1.0/dist/include
-- - LLVM Include path : D:/LLVM/19.1.0/dist/include
-- - Binary path : D:/LLVM/19.1.0/dist/bin
-- Found LLVM 19.1.1:
-- - CMake module path : D:/LLVM/19.1.1/dist/lib/cmake/llvm
-- - Clang include path : D:/LLVM/19.1.1/dist/include
-- - LLVM Include path : D:/LLVM/19.1.1/dist/include
-- - Binary path : D:/LLVM/19.1.1/dist/bin
-- ---- The below configuring for hipify-clang testing only ----
-- Found Python: C:/Users/TT/AppData/Local/Programs/Python/Python312/python.exe (found version "3.12.5") found components: Interpreter
-- Found Python: C:/Users/TT/AppData/Local/Programs/Python/Python312/python.exe (found version "3.12.7") found components: Interpreter
-- Found lit: C:/Users/TT/AppData/Local/Programs/Python/Python312/Scripts/lit.exe
-- Found FileCheck: D:/LLVM/19.1.0/dist/bin/FileCheck.exe
-- Found FileCheck: D:/LLVM/19.1.1/dist/bin/FileCheck.exe
-- Initial CUDA to configure:
-- - CUDA Toolkit path : C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.6
-- - CUDA Samples path : C:/ProgramData/NVIDIA Corporation/CUDA Samples/v12.5
-- - cuDNN path : D:/CUDA/cuDNN/9.4.0
-- - cuDNN path : D:/CUDA/cuDNN/9.5.0
-- - CUB path :
-- Found CUDAToolkit: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.6/include (found version "12.6.68")
-- Found CUDA config:
-- - CUDA Toolkit path : C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.6
-- - CUDA Samples path : C:/ProgramData/NVIDIA Corporation/CUDA Samples/v12.6
-- - cuDNN path : D:/CUDA/cuDNN/9.4.0
-- - CUDA Samples path : C:/ProgramData/NVIDIA Corporation/CUDA Samples/v12.5
-- - cuDNN path : D:/CUDA/cuDNN/9.5.0
-- - CUB path : C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v12.6/include/cub
-- Configuring done (1.4s)
-- Configuring done (2.1s)
-- Generating done (0.1s)
-- Build files have been written to: D:/HIPIFY/build
Expand Down
2 changes: 1 addition & 1 deletion docs/supported_apis.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@
| CUBLAS API | [HIP BLAS API](tables/CUBLAS_API_supported_by_HIP.md) | [ROC BLAS API](tables/CUBLAS_API_supported_by_ROC.md) | [HIP + ROC BLAS API](tables/CUBLAS_API_supported_by_HIP_and_ROC.md) |
| CUSPARSE API | [HIP SPARSE API](tables/CUSPARSE_API_supported_by_HIP.md) | [ROC SPARSE API](tables/CUSPARSE_API_supported_by_ROC.md) | [HIP + ROC SPARSE API](tables/CUSPARSE_API_supported_by_HIP_and_ROC.md) |
| CUSOLVER API | [HIP SOLVER API](tables/CUSOLVER_API_supported_by_HIP.md) | | |
| CURAND API | [HIP RAND API](tables/CURAND_API_supported_by_HIP.md) | | |
| CURAND API | [HIP RAND API](tables/CURAND_API_supported_by_HIP.md) |[ROC RAND API](tables/CURAND_API_supported_by_ROC.md) | [HIP + ROC RAND API](tables/CURAND_API_supported_by_HIP_and_ROC.md) |
| CUFFT API | [HIP FFT API](tables/CUFFT_API_supported_by_HIP.md) | | |
| CUDNN API | [HIP DNN API](tables/CUDNN_API_supported_by_HIP.md) | | |
| CUB API | [HIP CUB API](tables/CUB_API_supported_by_HIP.md) | | |
Expand Down
Loading

0 comments on commit f156b77

Please sign in to comment.