CMake updates #1181

victorapm · 2024-11-09T05:30:04Z

Finalized implementation of HIP support in CMake build
Achieve feature parity between CMake and Autotools builds
Upgrade CMake minimum required version to 3.21 (required by HIP)
Update CMake to use modern keywords and best practices
Better status messaging when running cmake
Add HYPRE_SetupGPUToolkit.cmake to manage GPU options in cmake build
Refactor top-level CMakeLists and moved helper functions to HYPRE_CMakeUtilities.cmake, e.g., setup_git_version_info for configuring and displaying hypre version using git version info
Add SOVERSION field to CMake build
Enable configuration of third-party libraries (TPLs) via CMake module
Add summary table of options used to configure hypre in CMake build
Remove HYPRE_INSTALL_PATH in favor of the commonly used CMAKE_INSTALL_PATH
Remove HYPRE_BUILD_TYPE in favor of the commonly used CMAKE_BUILD_TYPE
Remove HYPRE_ENABLE_SHARED in favor of the commonly used BUILD_SHARED_LIBS
Update documentation to reflect mainly CMake changes + a few fixes
- See https://hypre.readthedocs.io/en/cmake-hip/
Add CMake tests to machine-tioga
Add CTest support
Add HYPRE_ENABLE_LTO option to turn on link-time optimization

Closes #1104
Closes #673
Closes #1084
Closes #1073
Closes #1072
Closes #1039
Closes #907
Closes #767
Closes #757
Closes #771
Closes #501
Closes #473
Closes #228
Closes #928
Closes #508
Closes #555
Closes #502
Closes #464

…build structure

adam-sim-dev · 2024-11-28T01:20:27Z

cmake -B build -DHYPRE_ENABLE_CUDA=ON -DHYPRE_BUILD_EXAMPLES=ON -DHYPRE_ENABLE_UNIFIED_MEMORY=ON

-- Detecting CUDA GPU architectures using nvidia-smi...
CMake Warning at config/cmake/HYPRE_SetupCUDAToolkit.cmake:97 (message):
  C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6/bin/nvidia-smi.exe
  failed to execute:
Call Stack (most recent call first):
  config/cmake/HYPRE_SetupGPUToolkit.cmake:55 (include)
  CMakeLists.txt:358 (include)

This is strange. There is no nvidia-smi.exe in the folder C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6/bin. But after searching, there is one in C:\Windows\System32
and
C:\Windows\System32\DriverStore\FileRepository\nvle.inf_amd64_14964aa0497c7c40

See
https://stackoverflow.com/questions/57100015/how-do-i-run-nvidia-smi-on-windows

.gitignore

adam-sim-dev · 2024-11-28T01:53:17Z

I can build hypre with the following command line:

cmake -B build -DHYPRE_ENABLE_CUDA=ON -DHYPRE_BUILD_EXAMPLES=ON -DHYPRE_ENABLE_UNIFIED_MEMORY=ON -DCMAKE_CUDA_ARCHITECTURES=89

ex1 reports cuda error

>mpiexec -np 2 ./ex1

Running on "NVIDIA GeForce RTX 4060", major 8, minor 9, total memory 8.59 GB
MaxSharedMemoryPerBlock 49152, MaxSharedMemoryPerBlockOptin 101376
Running on "NVIDIA GeForce RTX 4060", major 8, minor 9, total memory 8.59 GB
MaxSharedMemoryPerBlock 49152, MaxSharedMemoryPerBlockOptin 101376
CUDA ERROR (code = 101, invalid device ordinal) at D:\hypre-main\hypre\src\utilities\memory.c:200
CUDA ERROR (code = 101, invalid device ordinal) at D:\hypre-main\hypre\src\utilities\memory.c:200
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:168
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:709
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:687
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:354
CUDA ERROR (code = 101, invalid device ordinal) at D:\hypre-main\hypre\src\utilities\memory.c:200
CUDA ERROR (code = 101, invalid device ordinal) at D:\hypre-main\hypre\src\utilities\memory.c:200
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:168
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:709
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:687
CUDA ERROR (code = 700, an illegal memory access was encountered) at D:\hypre-main\hypre\src\utilities\memory.c:354
Abort(-1) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, -1) - process 0
Abort(-1) on node 1 (rank 1 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, -1) - process 1

victorapm · 2024-11-28T05:16:13Z

@adam-sim-dev, this seems to be an issue with the NVIDIA driver. Could you check whether you are able to build and run the code: https://github.com/NVIDIA/cuda-samples/tree/master/Samples/1_Utilities/deviceQuery

adam-sim-dev · 2024-11-28T05:48:13Z

@adam-sim-dev, this seems to be an issue with the NVIDIA driver. Could you check whether you are able to build and run the code: https://github.com/NVIDIA/cuda-samples/tree/master/Samples/1_Utilities/deviceQuery

Test pass after I replace 12.5 with 12.6 in the .vcxproj:

CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "NVIDIA GeForce RTX 4060"
  CUDA Driver Version / Runtime Version          12.6 / 12.6
  CUDA Capability Major/Minor version number:    8.9
  Total amount of global memory:                 8188 MBytes (8585216000 bytes)
  (024) Multiprocessors, (128) CUDA Cores/MP:    3072 CUDA Cores
  GPU Max Clock rate:                            2460 MHz (2.46 GHz)
  Memory Clock rate:                             8501 Mhz
  Memory Bus Width:                              128-bit
  L2 Cache Size:                                 25165824 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total shared memory per multiprocessor:        102400 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1536
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 1 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  CUDA Device Driver Mode (TCC or WDDM):         WDDM (Windows Display Driver Model)
  Device supports Unified Addressing (UVA):      Yes
  Device supports Managed Memory:                Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      No
  Device PCI Domain ID / Bus ID / location ID:   0 / 1 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 12.6, CUDA Runtime Version = 12.6, NumDevs = 1
Result = PASS

victorapm · 2024-11-28T05:50:48Z

Ok, good! Could you also try running ex5 with on a single rank (-np 1)?

adam-sim-dev · 2024-11-28T05:59:09Z

Ok, good! Could you also try running ex5 with on a single rank (-np 1)?

Run in debug mode:

Running on "NVIDIA GeForce RTX 4060", major 8, minor 9, total memory 8.59 GB
MaxSharedMemoryPerBlock 49152, MaxSharedMemoryPerBlockOptin 101376
CUDA ERROR (code = 101, invalid device ordinal) at D:\hypre-main\hypre\src\utilities\memory.c:200
CUDA ERROR (code = 101, invalid device ordinal) at D:\hypre-main\hypre\src\utilities\memory.c:200
CUDA ERROR (code = 101, invalid device ordinal) at D:\hypre-main\hypre\src\utilities\memory.c:200
CUDA ERROR (code = 1, invalid argument) at D:\hypre-main\hypre\src\utilities\memory.c:662
CUDA error 1 [C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include\cub/util_device.cuh, 99]: invalid argument
CUDA error 101 [C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include\cub/util_device.cuh, 435]: invalid device ordinal
CUDA error 101 [C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.6\include\cub/device/dispatch/dispatch_scan.cuh, 557]: invalid device ordinal

victorapm · 2024-11-28T06:04:09Z

Thank you! It seems we will need to do more work on this. I'll let you know, thanks for helping

adam-sim-dev · 2024-11-28T06:08:46Z

Thank you! It seems we will need to do more work on this. I'll let you know, thanks for helping

My pleasure.

adam-sim-dev · 2024-11-28T07:03:07Z

Test on Ubuntu:

cmake -B cmbuild -DHYPRE_ENABLE_CUDA=ON -DHYPRE_ENABLE_UNIFIED_MEMORY=ON -DHYPRE_BUILD_EXAMPLES=ON

-- CMake Executable: /usr/bin/cmake
-- CMake Version: 3.30.2
-- Hypre Version: v2.32.0-69-gb678e6514 (cmake-hip)
-- The C compiler identification is GNU 11.4.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Build type: Release
-- Shared library: OFF
-- Installation directory: /home/Adam/Desktop/hypre/src/hypre
-- Using C standard: 99
-- Found MPI_C: /usr/lib/x86_64-linux-gnu/libmpich.so (found version "4.0")
-- Found MPI: TRUE (found version "4.0")
-- Adding MPI include directory: /usr/include/x86_64-linux-gnu/mpich
-- Performing Test HYPRE_HAVE_MPI_COMM_F2C
-- Performing Test HYPRE_HAVE_MPI_COMM_F2C - Success
-- The CXX compiler identification is GNU 11.4.0
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Enabling support for CXX.
-- Using CXX standard: C++14
-- Enabling CUDA toolkit
-- Using CUDA installation: /usr/local/cuda
-- The CUDA compiler identification is NVIDIA 12.4.131
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- Found CUDAToolkit: /usr/local/cuda/targets/x86_64-linux/include (found version "12.4.131")
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Detecting CUDA GPU architectures using nvidia-smi...
-- Detected CUDA GPU architectures: 89
-- CUDA Toolkit found at: /usr/local/cuda
-- CUDA Toolkit version: 12.4.131
-- CUDA Toolkit include directory: /usr/local/cuda/targets/x86_64-linux/include
-- CUDA Toolkit library directory: /usr/local/cuda/lib64
-- CUDA Thrust headers found in: /usr/local/cuda/targets/x86_64-linux/include
-- Found CUSPARSE library
-- Found CURAND library
-- Found CUBLAS library
-- Found CUBLASLT library
-- Found CUSOLVER library
-- Linking to CUDA libraries: CUDA::cusparse;CUDA::curand;CUDA::cublas;CUDA::cublasLt;CUDA::cusolver
-- CUDA C++ standard: 14
-- CUDA architectures: 89
-- CUDA flags:  --extended-lambda -ccbin=/usr/bin/c++
-- Using internal BLAS
-- Using internal LAPACK
-- Dependency directories: 
-- 
-- HYPRE Configuration Summary:
-- 
--  Base Options:
--  +-------------------------------------+---------+
--  | Option                              | Status  |
--  +-------------------------------------+---------+
--  | HYPRE_ENABLE_BIGINT                 | OFF     |
--  | HYPRE_ENABLE_MIXEDINT               | OFF     |
--  | HYPRE_ENABLE_SINGLE                 | OFF     |
--  | HYPRE_ENABLE_LONG_DOUBLE            | OFF     |
--  | HYPRE_ENABLE_COMPLEX                | OFF     |
--  | HYPRE_ENABLE_HYPRE_BLAS             | ON      |
--  | HYPRE_ENABLE_HYPRE_LAPACK           | ON      |
--  | HYPRE_ENABLE_PERSISTENT_COMM        | OFF     |
--  | HYPRE_ENABLE_FEI                    | OFF     |
--  | HYPRE_ENABLE_HOPSCOTCH              | OFF     |
--  | HYPRE_ENABLE_OPENMP                 | OFF     |
--  | HYPRE_ENABLE_MPI                    | ON      |
--  | HYPRE_ENABLE_PRINT_ERRORS           | OFF     |
--  | HYPRE_ENABLE_TEST_USING_HOST        | OFF     |
--  | HYPRE_ENABLE_HOST_MEMORY            | OFF     |
--  | HYPRE_BUILD_EXAMPLES                | ON      |
--  | HYPRE_BUILD_TESTS                   | OFF     |
--  +-------------------------------------+---------+
-- 
--  GPU Options:
--  +-------------------------------------+---------+
--  | Option                              | Status  |
--  +-------------------------------------+---------+
--  | HYPRE_ENABLE_CUDA                   | ON      |
--  | HYPRE_ENABLE_HIP                    | OFF     |
--  | HYPRE_ENABLE_SYCL                   | OFF     |
--  | HYPRE_ENABLE_GPU_AWARE_MPI          | OFF     |
--  | HYPRE_ENABLE_UNIFIED_MEMORY         | ON      |
--  | HYPRE_ENABLE_DEVICE_MALLOC_ASYNC    | OFF     |
--  | HYPRE_ENABLE_THRUST_ASYNC           | OFF     |
--  | HYPRE_ENABLE_GPU_PROFILING          | OFF     |
--  | HYPRE_ENABLE_CUDA_STREAMS           | ON      |
--  | HYPRE_ENABLE_CUSPARSE               | ON      |
--  | HYPRE_ENABLE_CUSOLVER               | ON      |
--  | HYPRE_ENABLE_CUBLAS                 | ON      |
--  | HYPRE_ENABLE_CURAND                 | ON      |
--  | HYPRE_ENABLE_DEVICE_POOL            | OFF     |
--  +-------------------------------------+---------+
-- 
--  Third-Party Library Options:
--  +-------------------------------------+---------+
--  | Option                              | Status  |
--  +-------------------------------------+---------+
--  | HYPRE_WITH_UMPIRE                   | OFF     |
--  | HYPRE_WITH_UMPIRE_HOST              | OFF     |
--  | HYPRE_WITH_UMPIRE_PINNED            | OFF     |
--  | HYPRE_WITH_UMPIRE_DEVICE            | OFF     |
--  | HYPRE_WITH_UMPIRE_UM                | OFF     |
--  | HYPRE_WITH_SUPERLU                  | OFF     |
--  | HYPRE_WITH_DSUPERLU                 | OFF     |
--  | HYPRE_WITH_MAGMA                    | OFF     |
--  | HYPRE_WITH_CALIPER                  | OFF     |
--  +-------------------------------------+---------+
-- 
-- Configuring done (1.2s)
-- Generating done (0.1s)
-- Build files have been written to: /home/Adam/Desktop/hypre/src/cmbuild

Hypre library could be built successfully. However, examples build failed.

make ex1

[ 98%] Built target HYPRE
[ 98%] Building CUDA object examples/CMakeFiles/ex1.dir/ex1.c.o
[100%] Linking CXX executable ex1
lto-wrapper: warning: using serial compilation of 6 LTRANS jobs
/tmp/ccUh00Pj.s: Assembler messages:
/tmp/ccUh00Pj.s:49: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:254: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:459: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:1106: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:2432: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:6418: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:8054: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:8499: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:128772: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:128977: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:129020: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:129383: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:129871: Error: symbol `fatbinData' is already defined
/tmp/ccUh00Pj.s:130633: Error: symbol `fatbinData' is already defined
lto-wrapper: fatal error: /usr/bin/c++ returned 1 exit status
compilation terminated.
/usr/bin/ld: error: lto-wrapper failed
collect2: error: ld returned 1 exit status
make[3]: *** [examples/CMakeFiles/ex1.dir/build.make:107: examples/ex1] Error 1
make[2]: *** [CMakeFiles/Makefile2:587: examples/CMakeFiles/ex1.dir/all] Error 2
make[1]: *** [CMakeFiles/Makefile2:594: examples/CMakeFiles/ex1.dir/rule] Error 2
make: *** [Makefile:208: ex1] Error 2

victorapm · 2024-11-28T18:01:04Z

Hi Adam, I can't reproduce this on Ubuntu, gcc-11 and nvcc 12.6. Could you try a fresh build (remove cmbuild and call cmake again). Another idea is to disable LTO and enable verbose mode with -DCMAKE_INTERPROCEDURAL_OPTIMIZATION=OFF -DCMAKE_VERBOSE_MAKEFILE=ON

adam-sim-dev · 2024-11-29T01:02:29Z

Hi Adam, I can't reproduce this on Ubuntu, gcc-11 and nvcc 12.6. Could you try a fresh build (remove cmbuild and call cmake again). Another idea is to disable LTO and enable verbose mode with -DCMAKE_INTERPROCEDURAL_OPTIMIZATION=OFF -DCMAKE_VERBOSE_MAKEFILE=ON

ex1.log
cmake.log
HYPRE.log

victorapm · 2024-11-29T04:27:20Z

Adam, did you use -DCMAKE_INTERPROCEDURAL_OPTIMIZATION=OFF? Can you share CMakeCache.txt?

adam-sim-dev · 2024-11-29T04:53:22Z

-DCMAKE_INTERPROCEDURAL_OPTIMIZATION=OFF -DCMAKE_VERBOSE_MAKEFILE=ON

Yes, I copied -DCMAKE_INTERPROCEDURAL_OPTIMIZATION=OFF -DCMAKE_VERBOSE_MAKEFILE=ON

CMakeCache.txt

victorapm · 2024-12-02T14:45:10Z

Thanks! Now I see your MPI installation is exporting LTO-related compile flags to hypre, which causes issues with the device build. I've added a check to remove those flags when they come from the MPI dependency and also a new option called HYPRE_ENABLE_LTO to turn on link-time optimization in hypre (off by default).

rfalgout · 2024-12-09T18:37:13Z

src/docs/usr-manual/ch-misc.rst

-systems, such as Linux, AIX, and Mac OS X.  Then in `CMake instructions`_, we
-explain an alternative approach using the CMake build system [CMakeWeb]_, which
-is the best approach for building hypre on Windows systems in particular.
+After obtaining the source code (see :ref:`getting-source`), there are two main ways to build hypre:


three main ways

rfalgout · 2024-12-09T18:39:09Z

src/docs/usr-manual/ch-misc.rst


-Hypre can support NVIDIA GPUs with CUDA and OpenMP (:math:`{\ge}` 4.5). The related ``configure`` options are
+[Spack_](https://spack.io/) is a package manager designed for supercomputers, Linux, and macOS.


There is a rendering issue here with the spack link

rfalgout · 2024-12-09T18:42:42Z

src/docs/usr-manual/ch-misc.rst

+GPU Build Options
+==============================================================
+
+hypre provides support for multiple GPU architectures through different programming models:


Don't start a sentence with "hypre". Change to "The hypre library".

rfalgout · 2024-12-09T18:45:46Z

src/docs/usr-manual/ch-misc.rst


-.. _CMake options:
+hypre provides CMake configuration files that enable easy integration. Create a


The hypre library

rfalgout · 2024-12-09T18:50:05Z

src/docs/usr-manual/ch-sstruct.rst


 As with the ``Struct`` interface, each process describes that portion of the
-grid that it "owns", one box at a time.  Figure :ref:`fig-sstruct-grid` shows
+grid that it "owns", one box at a time. `Figure 8 <fig-sstruct-grid>`_ shows


This link still doesn't work

rfalgout · 2024-12-09T18:52:06Z

src/docs/usr-manual/ch-sstruct.rst

@@ -242,8 +247,8 @@ describes stencil entry 5 in the figure is given the entry number 5, the offset
 to which this entry couples).  Recall from Figure :ref:`fig-gridvars` the
 convention used for referencing variables of different types.  The geometry
 description uses the same convention, but with indices numbered relative to the
-referencing index :math:`(0,0)` for the stencil's center.  Figure
-:ref:`fig-sstruct-graph` shows the code for setting up the graph .
+referencing index :math:`(0,0)` for the stencil's center. `Figure 9 <fig-sstruct-graph>`_


This link doesn't work either

rfalgout · 2024-12-09T18:52:40Z

src/docs/usr-manual/ch-sstruct.rst

@@ -253,7 +258,7 @@ referencing index :math:`(0,0)` for the stencil's center.  Figure
   Figure 7a


This figure label no longer makes since with the addition of Figs 8 and 9

rfalgout · 2024-12-09T18:53:29Z

src/docs/usr-manual/ch-sstruct.rst

-(the code for the other processes is similar).  The "icons" at the top of the
-figure illustrate the result of the numbered lines of code.  Process 0 needs to
-describe the data pictured in the bottom-right of the figure.  That is, it needs
+portion of the grid that it "owns", one box at a time. `Figure 11 <fig-sstruct-fem-grid>`_


Link doesn't work

rfalgout

This looks good to me, except that the manual still has some figure issues. If you find a solution, I'd like to see it before you merge. Otherwise, looks great! Thanks!

victorapm and others added 30 commits September 12, 2024 22:51

Add HYPRE_Setup*Toolkit.cmake files

46916f3

Add HYPRE_USING_GPU to define C++ source files

7583390

Initial working version for HIP support

d79a204

Add summary table + misc improvements

777875d

Change to modern CMake practice

8690017

Remove unused variable

31b0d97

Add configure_mpi_target function

f7c411f

Modernizing CUDA cmake build

25491ae

Add missing HYPRE_CUDA/HIP_CALL macros

1fda469

Allow CUDA/HIP builds with specific toolkit locations

f936f14

Improve CUDA and HIP builds robustness

ee9f9d5

Adjustment to run nvidia-smi to detect cuda arch

f4d3f4d

Simplify build options setup with set_conditional_var + Simplify GPU …

329729f

…build structure

Add setup_tpl to simplify TPLs usage

c7da232

Add find_package support to setup_tpl

c1f9961

Add VERSION and SOVERSION info

3859516

Modernize shared lib build: improve dependency linkage and CMake setup

5b8a310

Top-level CMakeLists refactoring

6e08c7d

Simplify build with set_hypre_option

fcb58e6

Add ensure_options_match/differ macros

14f58d8

Fix target: CUDAToolkit -> CUDA

657d0d5

Improve dependency search in HYPREConfig.cmake

34cd9cd

Adjustments

9212a93

Merge branch 'master' into cmake-hip

c032f71

Merge branch 'master' into cmake-hip

254f586

Merge branch 'master' into cmake-hip

0b79d5a

Update SOVERSION for consistency with ABI approach in hypre

f60f359

Merge branch 'master' into cmake-hip

1c9fb99

cmake: update soversion (#1164)

35a8c91

Merge branch 'master' into cmake-hip

b4a5fbe

adam-sim-dev reviewed Nov 28, 2024

View reviewed changes

.gitignore Show resolved Hide resolved

victorapm added 3 commits November 27, 2024 21:30

Fix nvidia-smi search on Windows

d6bc5ee

Ignore Windows executables

65cd3d5

Update cmake variables that are really internal

b678e65

Fix LTO-related issues and add HYPRE_ENABLE_LTO build option

a44dfde

victorapm mentioned this pull request Dec 2, 2024

Setting HIP compiler #555

Open

Updating cmake.sh autotest script to work with cmake build changes

9649c9a

rfalgout reviewed Dec 9, 2024

View reviewed changes

rfalgout approved these changes Dec 9, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

CMake updates #1181

CMake updates #1181

victorapm commented Nov 9, 2024 •

edited

Loading

adam-sim-dev commented Nov 28, 2024 •

edited

Loading

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 29, 2024

victorapm commented Nov 29, 2024

adam-sim-dev commented Nov 29, 2024

victorapm commented Dec 2, 2024

rfalgout Dec 9, 2024

rfalgout Dec 9, 2024

rfalgout Dec 9, 2024

rfalgout Dec 9, 2024

rfalgout Dec 9, 2024

rfalgout Dec 9, 2024

rfalgout Dec 9, 2024

rfalgout Dec 9, 2024

rfalgout left a comment


		Hypre can support NVIDIA GPUs with CUDA and OpenMP (:math:`{\ge}` 4.5). The related ``configure`` options are
		[Spack_](https://spack.io/) is a package manager designed for supercomputers, Linux, and macOS.


		.. _CMake options:
		hypre provides CMake configuration files that enable easy integration. Create a

		@@ -253,7 +258,7 @@ referencing index :math:`(0,0)` for the stencil's center. Figure
		Figure 7a

CMake updates #1181

Are you sure you want to change the base?

CMake updates #1181

Conversation

victorapm commented Nov 9, 2024 • edited Loading

adam-sim-dev commented Nov 28, 2024 • edited Loading

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

adam-sim-dev commented Nov 28, 2024

victorapm commented Nov 28, 2024

adam-sim-dev commented Nov 29, 2024

victorapm commented Nov 29, 2024

adam-sim-dev commented Nov 29, 2024

victorapm commented Dec 2, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rfalgout left a comment

Choose a reason for hiding this comment

victorapm commented Nov 9, 2024 •

edited

Loading

adam-sim-dev commented Nov 28, 2024 •

edited

Loading