diff options
Diffstat (limited to 'dev-util/Tensile')
20 files changed, 552 insertions, 417 deletions
diff --git a/dev-util/Tensile/Manifest b/dev-util/Tensile/Manifest index c3e8f48bab20..4d93ffa204a1 100644 --- a/dev-util/Tensile/Manifest +++ b/dev-util/Tensile/Manifest @@ -1,3 +1,4 @@ DIST Tensile-5.0.2-PR1419.patch.gz 43754 BLAKE2B 3b1d7333ed0115f77816806e7735e25d897dbae3f7a40ece61d655c8f016d9d1c804b586b4e0e75661e20c517f5937a61d5ead25a545f0ce2e4a178bcbb3e7be SHA512 0154d8e14d335f4ecd86e5db335289f8662cede7e5d4b93bf6c56d8e48c1d7d0457271c078c9abbbca1f4aa91f028f6ce6a89efb783ba0207d0f39fd5de8a42c -DIST rocm-Tensile-4.3.0.tar.gz 14250149 BLAKE2B 030138eaca2a0aadd96801e6dbd72e510716dd90553ef3795c5e04e00a34a05ecae82b24f755e4033a4acfcdb1cf26291da1e7902bb090f89a010d403e832beb SHA512 126db0b413c716fba8d5be9bff7a44fd1badacbf32f3db8d0db649819177db37ebd56fd22dd3c809655f5d29675be115e698cd10bc3d0b4b23878ae3726fce47 -DIST rocm-Tensile-5.0.2.tar.gz 14362090 BLAKE2B d44c3135496db41c4eb2545f6054c84493e89ef6357da881dae6859763338417e7a9e142cdce7e2de0536684bb641049ad0440964adbe68d03ca329f01efabc9 SHA512 f5ab7d7cbc5ba25498a02a6b7c07d5a9a548edcd44b439b92c8718072382adc18ae78274db1a7174e06436ac1396f3b4edafc7f491a2c87086c368e7270e64e3 +DIST rocm-Tensile-5.1.3.tar.gz 14357517 BLAKE2B 7c5881cc2b1d25db39f54a99f9f0a08395a5f7a58deebe74d9868e8f7044db5ae2f3907196adf16dc6abeba0f1f47f5cce5c8900891276daff8e4ad614cf3b94 SHA512 4405cb6de65c1da1fd93910fc901402a24e996ff04eccadf483c25d8eb518e03ea9bbf00e372dd5fe377696e71c39355603095f3c5fc1e26d311ed15d8e3b734 +DIST rocm-Tensile-5.4.2.tar.gz 14413805 BLAKE2B ca1c13785164439f6fe83f16b36b7a0b06a63b70b9f0741d77b0fff98f38d1e1c7e9a54ef61897e6eb0a4eaf52ec9e5f14f42a7e6fd657f26306afa3de0c74a2 SHA512 d99736a202c7353979e732941199b4b20e53ff2b6c19a7175609440a840aa5624bc53a8013c435f31ac37933527771521ff02137c1c8fd53c5317eca8533be80 +DIST rocm-Tensile-5.7.1.tar.gz 12396002 BLAKE2B ec2c2bc3f07fe22abd800b642c998ea80b9e4e2d0d28c4763d574dbd2fcb14c735223ddc7571d4ab3a48f0bd819cc835c16df9882049c5bf56f25ef729e6d654 SHA512 8f356f36cfec74ee9b000d1180cf52077dc5c595e5c5a40792fc3eb13f3139980c7440f9bc912218f946cf8137e2481d23fe39f110511c1bb55fe85a933ffd36 diff --git a/dev-util/Tensile/Tensile-5.0.2-r2.ebuild b/dev-util/Tensile/Tensile-5.1.3-r3.ebuild index 38854fd0c269..da1adcab4f7b 100644 --- a/dev-util/Tensile/Tensile-5.0.2-r2.ebuild +++ b/dev-util/Tensile/Tensile-5.1.3-r3.ebuild @@ -1,16 +1,18 @@ -# Copyright 1999-2022 Gentoo Authors +# Copyright 1999-2023 Gentoo Authors # Distributed under the terms of the GNU General Public License v2 EAPI=8 -PYTHON_COMPAT=( python3_{8..10} ) +PYTHON_COMPAT=( python3_{9..11} ) DISTUTILS_USE_PEP517=setuptools -inherit distutils-r1 prefix +inherit distutils-r1 llvm prefix + +LLVM_MAX_SLOT=15 DESCRIPTION="Stretching GPU performance for GEMMs and tensor contractions" HOMEPAGE="https://github.com/ROCmSoftwarePlatform/Tensile" SRC_URI="https://github.com/ROCmSoftwarePlatform/Tensile/archive/rocm-${PV}.tar.gz -> rocm-Tensile-${PV}.tar.gz - https://github.com/littlewu2508/littlewu2508.github.io/raw/main/gentoo-distfiles/${P}-PR1419.patch.gz" + https://github.com/littlewu2508/littlewu2508.github.io/raw/main/gentoo-distfiles/${PN}-5.0.2-PR1419.patch.gz" S="${WORKDIR}/${PN}-rocm-${PV}" LICENSE="MIT" @@ -21,27 +23,29 @@ SLOT="0/$(ver_cut 1-2)" RESTRICT="test" RDEPEND="${PYTHON_DEPS} + >=dev-cpp/msgpack-cxx-6.0.0 dev-python/pyyaml[${PYTHON_USEDEP}] dev-python/msgpack[${PYTHON_USEDEP}] + dev-util/hip + sys-devel/clang:${LLVM_MAX_SLOT} >=dev-util/rocm-smi-4.3.0 " -DEPEND="${RDEPEND} - dev-util/hip:= -" +DEPEND="${RDEPEND}" -PATCHES=( "${FILESDIR}"/${PN}-4.3.0-output-commands.patch - "${FILESDIR}"/${PN}-5.0.2-gfx1031.patch - "${FILESDIR}"/${PN}-5.0.2-fix-arch-parse.patch - "${FILESDIR}"/${PN}-5.0.2-use-ninja.patch - "${FILESDIR}"/${PN}-5.0.2-adjust-parallel.patch - "${WORKDIR}"/${PN}-5.0.2-PR1419.patch - ) +PATCHES=( + "${FILESDIR}"/${PN}-change-cmake-name-for-msgpack-cxx-6-release.patch + "${FILESDIR}"/${PN}-4.3.0-output-commands.patch + "${FILESDIR}"/${PN}-5.0.2-gfx1031.patch + "${FILESDIR}"/${PN}-5.0.2-fix-arch-parse.patch + "${FILESDIR}"/${PN}-5.0.2-use-ninja.patch +) CMAKE_USE_DIR="${WORKDIR}/Source" src_prepare() { distutils-r1_src_prepare - eapply $(prefixify_ro "${FILESDIR}"/${PN}-5.0.2-gentoopath.patch) + sed -e "s,\@LLVM_PATH\@,$(get_llvm_prefix ${LLVM_MAX_SLOT}),g" "${FILESDIR}"/${PN}-5.1.3-gentoopath.patch > "${S}"/gentoopath.patch || die + eapply $(prefixify_ro "${S}"/gentoopath.patch) pushd ${PN} || die diff --git a/dev-util/Tensile/Tensile-4.3.0-r1.ebuild b/dev-util/Tensile/Tensile-5.4.2-r2.ebuild index 7cb7e94cd256..c846dce60997 100644 --- a/dev-util/Tensile/Tensile-4.3.0-r1.ebuild +++ b/dev-util/Tensile/Tensile-5.4.2-r2.ebuild @@ -1,11 +1,14 @@ -# Copyright 1999-2022 Gentoo Authors +# Copyright 1999-2023 Gentoo Authors # Distributed under the terms of the GNU General Public License v2 EAPI=8 -PYTHON_COMPAT=( python3_{8..10} ) +PYTHON_COMPAT=( python3_{10..11} ) DISTUTILS_USE_PEP517=setuptools -inherit distutils-r1 prefix +ROCM_VERSION=${PV} +inherit cmake distutils-r1 llvm prefix rocm + +LLVM_MAX_SLOT=15 DESCRIPTION="Stretching GPU performance for GEMMs and tensor contractions" HOMEPAGE="https://github.com/ROCmSoftwarePlatform/Tensile" @@ -15,29 +18,37 @@ S="${WORKDIR}/${PN}-rocm-${PV}" LICENSE="MIT" KEYWORDS="~amd64" SLOT="0/$(ver_cut 1-2)" +IUSE="client" +REQUIRED_USE="client? ( ${ROCM_REQUIRED_USE} )" # Not compatible with recent versions of pytest RESTRICT="test" RDEPEND="${PYTHON_DEPS} + >=dev-cpp/msgpack-cxx-6.0.0 dev-python/pyyaml[${PYTHON_USEDEP}] dev-python/msgpack[${PYTHON_USEDEP}] + dev-util/hip >=dev-util/rocm-smi-4.3.0 + sys-devel/clang:${LLVM_MAX_SLOT} " -DEPEND="${RDEPEND} - dev-util/hip:${SLOT} -" +DEPEND="${RDEPEND}" + +PATCHES=( + "${FILESDIR}"/${PN}-change-cmake-name-for-msgpack-cxx-6-release.patch + "${FILESDIR}"/${PN}-4.3.0-output-commands.patch + "${FILESDIR}"/${PN}-5.4.2-gfx1031.patch + "${FILESDIR}"/${PN}-5.4.2-fix-arch-parse.patch + "${FILESDIR}"/${PN}-5.4.2-use-ninja.patch +) -PATCHES=( "${FILESDIR}"/${PN}-4.3.0-output-commands.patch - "${FILESDIR}"/${PN}-4.3.0-hsaco-compile-specified-arch.patch - "${FILESDIR}"/${PN}-4.3.0-gfx1031.patch - "${FILESDIR}"/${PN}-4.3.0-fix-arch-parse.patch - "${FILESDIR}"/${PN}-4.3.0-use-ninja.patch - ) +CMAKE_USE_DIR="${S}/${PN}/Source" src_prepare() { distutils-r1_src_prepare - eapply $(prefixify_ro "${FILESDIR}"/${PN}-4.3.0-gentoopath.patch) + sed -e "s,\@LLVM_PATH\@,$(get_llvm_prefix ${LLVM_MAX_SLOT}),g" \ + "${FILESDIR}"/${PN}-5.4.2-gentoopath.patch > "${S}"/gentoopath.patch || die + eapply $(prefixify_ro "${S}"/gentoopath.patch) pushd ${PN} || die @@ -46,24 +57,41 @@ src_prepare() { sed -r -e "/TENSILE_USE_LLVM/s/ON/OFF/" \ -i Source/CMakeLists.txt || die sed -e "/chmod 755/d" -i Source/TensileCreateLibrary.cmake || die # remove chmod 755 on - sed -e "s,\${Tensile_ROOT}/bin/,,g" -i Source/TensileCreateLibrary.cmake cmake/TensileConfig.cmake || die # ${Tensile_ROOT}/bin does not exists; call command directly - local Tensile_share_dir="\"${EPREFIX}/usr/share/${PN}\"" - sed -e "/HipClangVersion/s/0,0,0/$(hipconfig -v)/" \ - -e "/SourcePath/s,globalParameters\[\"ScriptPath\"\],${Tensile_share_dir}," \ - -i Common.py || die - - sed -e "/CMAKE_CXX_COMPILER/s,globalParameters\[\"ROCmBinPath\"\],\"${EPREFIX}/usr/lib/hip/bin\"," -i ClientExecutable.py || die + # ${Tensile_ROOT}/bin does not exists; call command directly + sed -e "s,\${Tensile_ROOT}/bin/,,g" -i Source/TensileCreateLibrary.cmake cmake/TensileConfig.cmake || die - sed -e "/scriptDir/s,os.path.dirname(os.path.realpath(__file__)),${Tensile_share_dir}," -i ReplacementKernels.py || die + local Tensile_share_dir="\"${EPREFIX}/usr/share/${PN}\"" + sed -e "/HipClangVersion/s/0.0.0/$(hipconfig -v)/" -i Common.py || die - sed -e "s,os.path.dirname(os.path.realpath(__file__)),${Tensile_share_dir},g" -i ${PN}.py || die + sed -e "s,os.path.dirname(os.path.realpath(__file__)),${Tensile_share_dir},g" \ + -i ReplacementKernels.py Common.py ${PN}.py || die sed -e "s|os\.path\.dirname.*$|\"${EPREFIX}/usr/share/Tensile/Source\", end='')|" -i __init__.py || die popd || die sed -e "/package_data/d" -e "/data_files/d" -i setup.py || die + use client && PATCHES= cmake_src_prepare # do not apply patches again in cmake_src_prepare +} + +src_configure() { + distutils-r1_src_configure + if use client; then + local mycmakeargs=( + -DCMAKE_SKIP_RPATH=ON + -DTENSILE_USE_MSGPACK=ON + -DTENSILE_USE_LLVM=ON + -DTensile_LIBRARY_FORMAT=msgpack + -DAMDGPU_TARGETS="$(get_amdgpu_flags)" + ) + CXX=hipcc cmake_src_configure + fi +} + +src_compile() { + distutils-r1_src_compile + use client && cmake_src_compile } python_install() { @@ -80,7 +108,12 @@ src_install() { pushd ${PN} || die insinto /usr/share/${PN} - doins -r Configs Perf ReplacementKernels ReplacementKernels-cov3 Source + doins -r Configs Perf ReplacementKernels ReplacementKernels-cov3 Source CustomKernels insinto /usr/$(get_libdir)/cmake/${PN} doins cmake/*.cmake + + if use client; then + pushd "${BUILD_DIR}" || die + dobin client/tensile_client + fi } diff --git a/dev-util/Tensile/Tensile-5.7.1-r2.ebuild b/dev-util/Tensile/Tensile-5.7.1-r2.ebuild new file mode 100644 index 000000000000..94b4aac75b1b --- /dev/null +++ b/dev-util/Tensile/Tensile-5.7.1-r2.ebuild @@ -0,0 +1,133 @@ +# Copyright 1999-2024 Gentoo Authors +# Distributed under the terms of the GNU General Public License v2 + +EAPI=8 + +PYTHON_COMPAT=( python3_{10..12} ) +DISTUTILS_USE_PEP517=setuptools +ROCM_VERSION=${PV} +inherit cmake distutils-r1 llvm prefix rocm + +LLVM_MAX_SLOT=17 + +DESCRIPTION="Stretching GPU performance for GEMMs and tensor contractions" +HOMEPAGE="https://github.com/ROCmSoftwarePlatform/Tensile" +SRC_URI="https://github.com/ROCmSoftwarePlatform/Tensile/archive/rocm-${PV}.tar.gz -> rocm-Tensile-${PV}.tar.gz" +S="${WORKDIR}/${PN}-rocm-${PV}" + +LICENSE="MIT" +KEYWORDS="~amd64" +SLOT="0/$(ver_cut 1-2)" +IUSE="client test" +REQUIRED_USE="client? ( ${ROCM_REQUIRED_USE} )" + +RESTRICT="!test? ( test )" + +RDEPEND="${PYTHON_DEPS} + client? ( dev-libs/boost ) + >=dev-cpp/msgpack-cxx-6.0.0 + dev-python/pyyaml[${PYTHON_USEDEP}] + dev-python/msgpack[${PYTHON_USEDEP}] + =dev-util/hip-5* + >=dev-util/rocm-smi-4.3.0 + sys-devel/clang:${LLVM_MAX_SLOT} +" +DEPEND="${RDEPEND}" +BDEPEND=" + test? ( + dev-python/pytest-forked[${PYTHON_USEDEP}] + dev-python/pytest-xdist[${PYTHON_USEDEP}] + dev-python/filelock[${PYTHON_USEDEP}] + ) +" + +distutils_enable_tests pytest + +PATCHES=( + "${FILESDIR}"/${PN}-4.3.0-output-commands.patch + "${FILESDIR}"/${PN}-5.4.2-fix-arch-parse.patch + "${FILESDIR}"/${PN}-5.4.2-use-ninja.patch + "${FILESDIR}"/${PN}-5.7.1-expand-isa-compatibility.patch + "${FILESDIR}"/${PN}-5.7.1-fix-msgpack-dependency.patch +) + +CMAKE_USE_DIR="${S}/${PN}/Source" + +src_prepare() { + distutils-r1_src_prepare + sed -e "s,\@LLVM_PATH\@,$(get_llvm_prefix ${LLVM_MAX_SLOT}),g" \ + "${FILESDIR}"/${PN}-5.7.1-gentoopath.patch > "${S}"/gentoopath.patch || die + eapply $(prefixify_ro "${S}"/gentoopath.patch) + + pushd ${PN} || die + + sed -e "/ROCM_SMI_ROOT/s,lib,$(get_libdir)," \ + -i Source/cmake/FindROCmSMI.cmake || die + sed -r -e "/TENSILE_USE_LLVM/s/ON/OFF/" \ + -i Source/CMakeLists.txt || die + + # ${Tensile_ROOT}/bin does not exists; call command directly + sed -e "s,\${Tensile_ROOT}/bin/,,g" -i cmake/TensileConfig.cmake || die + + local Tensile_share_dir="\"${EPREFIX}/usr/share/${PN}\"" + sed -e "/HipClangVersion/s/0.0.0/$(hipconfig -v)/" -i Common.py || die + + sed -e "s,os.path.dirname(os.path.realpath(__file__)),${Tensile_share_dir},g" \ + -i ReplacementKernels.py Common.py ${PN}.py || die + + sed -e "s|os\.path\.dirname.*$|\"${EPREFIX}/usr/share/Tensile/Source\", end='')|" -i __init__.py || die + + popd || die + + sed -e "/package_data/d" -e "/data_files/d" -i setup.py || die + use client && PATCHES= cmake_src_prepare # do not apply patches again in cmake_src_prepare +} + +src_configure() { + distutils-r1_src_configure + if use client; then + local mycmakeargs=( + -DCMAKE_SKIP_RPATH=ON + -DTENSILE_USE_MSGPACK=ON + -DTENSILE_USE_LLVM=ON + -DTensile_LIBRARY_FORMAT=msgpack + -DAMDGPU_TARGETS="$(get_amdgpu_flags)" + ) + CXX=hipcc cmake_src_configure + fi +} + +src_compile() { + distutils-r1_src_compile + use client && cmake_src_compile +} + +python_install() { + distutils-r1_python_install + + python_moduleinto Tensile + pushd Tensile || die + python_domodule Components + python_newexe Utilities/merge.py ${PN}-merge +} + +src_install() { + distutils-r1_src_install + + pushd ${PN} || die + insinto /usr/share/${PN} + doins -r Configs Perf ReplacementKernels-cov3 Source CustomKernels + insinto /usr/$(get_libdir)/cmake/${PN} + doins cmake/*.cmake + + if use client; then + pushd "${BUILD_DIR}" || die + dobin client/tensile_client + fi +} + +# Test suite fails to start without this +python_test() { + export ROCM_PATH="${EPREFIX}/usr" + epytest +} diff --git a/dev-util/Tensile/files/Tensile-4.3.0-fix-arch-parse.patch b/dev-util/Tensile/files/Tensile-4.3.0-fix-arch-parse.patch deleted file mode 100644 index 01b94b32cb6e..000000000000 --- a/dev-util/Tensile/files/Tensile-4.3.0-fix-arch-parse.patch +++ /dev/null @@ -1,12 +0,0 @@ -Index: Tensile-rocm-4.3.0/Tensile/Tensile.py -=================================================================== ---- Tensile-rocm-4.3.0.orig/Tensile/Tensile.py -+++ Tensile-rocm-4.3.0/Tensile/Tensile.py -@@ -97,7 +97,6 @@ def addCommonArguments(argParser): - """ - - (key, value) = par.split("=") -- value = eval(value) - return (key, value) - - argParser.add_argument("-d", "--device", dest="device", type=int, \ diff --git a/dev-util/Tensile/files/Tensile-4.3.0-gentoopath.patch b/dev-util/Tensile/files/Tensile-4.3.0-gentoopath.patch deleted file mode 100644 index da135e3b38ee..000000000000 --- a/dev-util/Tensile/files/Tensile-4.3.0-gentoopath.patch +++ /dev/null @@ -1,62 +0,0 @@ -Index: Tensile-rocm-4.3.0/Tensile/Source/client/CMakeLists.txt -=================================================================== ---- Tensile-rocm-4.3.0.orig/Tensile/Source/client/CMakeLists.txt -+++ Tensile-rocm-4.3.0/Tensile/Source/client/CMakeLists.txt -@@ -50,3 +50,5 @@ endif() - foreach(arch IN LISTS TENSILE_GPU_ARCHS) - target_link_libraries(tensile_client PRIVATE "--amdgpu-target=${arch}") - endforeach(arch) -+ -+target_link_libraries(tensile_client PRIVATE LLVMSupport) -Index: Tensile-rocm-4.3.0/Tensile/Source/lib/CMakeLists.txt -=================================================================== ---- Tensile-rocm-4.3.0.orig/Tensile/Source/lib/CMakeLists.txt -+++ Tensile-rocm-4.3.0/Tensile/Source/lib/CMakeLists.txt -@@ -48,7 +48,7 @@ set(tensile_sources ${tensile_sources} - ) - - if(TENSILE_USE_LLVM) -- find_package(LLVM 13.0 QUIET CONFIG) -+ find_package(LLVM PATHS @GENTOO_PORTAGE_EPREFIX@/usr/lib/llvm/roc QUIET CONFIG) - if(NOT LLVM_FOUND) - find_package(LLVM 12.0 QUIET CONFIG) - if(NOT LLVM_FOUND) -Index: Tensile-rocm-4.3.0/Tensile/Common.py -=================================================================== ---- Tensile-rocm-4.3.0.orig/Tensile/Common.py -+++ Tensile-rocm-4.3.0/Tensile/Common.py -@@ -1667,7 +1667,7 @@ def assignGlobalParameters( config ): - else: - print2(" %24s: %8s (unspecified)" % (key, defaultValue)) - -- globalParameters["ROCmPath"] = "/opt/rocm" -+ globalParameters["ROCmPath"] = "@GENTOO_PORTAGE_EPREFIX@/usr" - if "ROCM_PATH" in os.environ: - globalParameters["ROCmPath"] = os.environ.get("ROCM_PATH") - if "TENSILE_ROCM_PATH" in os.environ: -Index: Tensile-rocm-4.3.0/Tensile/ClientWriter.py -=================================================================== ---- Tensile-rocm-4.3.0.orig/Tensile/ClientWriter.py -+++ Tensile-rocm-4.3.0/Tensile/ClientWriter.py -@@ -284,7 +284,7 @@ def getBuildNewClientLibraryScript(build - import io - runScriptFile = io.StringIO() - -- callCreateLibraryCmd = globalParameters["ScriptPath"] + "/bin/TensileCreateLibrary" -+ callCreateLibraryCmd = "TensileCreateLibrary" - - - if globalParameters["MergeFiles"]: -Index: Tensile-rocm-4.3.0/Tensile/GenerateSummations.py -=================================================================== ---- Tensile-rocm-4.3.0.orig/Tensile/GenerateSummations.py -+++ Tensile-rocm-4.3.0/Tensile/GenerateSummations.py -@@ -57,7 +57,7 @@ def createLibraryForBenchmark(logicPath, - Selection. - """ - -- pythonExePath = os.path.join(os.path.dirname(os.path.realpath(__file__)), "bin", "TensileCreateLibrary") -+ pythonExePath = "TensileCreateLibrary" - args = [pythonExePath, \ - "--merge-files", "--no-legacy-components", \ - "--new-client-only", "--no-short-file-names", "--no-library-print-debug", "--architecture=all", \ diff --git a/dev-util/Tensile/files/Tensile-4.3.0-gfx1031.patch b/dev-util/Tensile/files/Tensile-4.3.0-gfx1031.patch deleted file mode 100644 index 99abceb7c44d..000000000000 --- a/dev-util/Tensile/files/Tensile-4.3.0-gfx1031.patch +++ /dev/null @@ -1,119 +0,0 @@ ---- Tensile-rocm-4.3.0/pytest.ini 2021-04-22 22:35:33.000000000 +0800 -+++ Tensile-rocm-4.3.0/pytest.ini 2021-09-30 17:32:21.382806851 +0800 -@@ -88,6 +88,7 @@ - xfail-gfx1011: architecture - xfail-gfx1012: architecture - xfail-gfx1030: architecture -+ xfail-gfx1031: architecture - skip-gfx000: architecture - skip-gfx900: architecture - skip-gfx906: architecture -@@ -97,3 +98,4 @@ - skip-gfx1011: architecture - skip-gfx1012: architecture - skip-gfx1030: architecture -+ skip-gfx1031: architecture ---- Tensile-rocm-4.3.0/Tensile/Common.py 2021-10-05 14:41:22.490754124 +0800 -+++ Tensile-rocm-4.3.0/Tensile/Common.py 2021-09-30 17:33:19.041724015 +0800 -@@ -200,7 +200,7 @@ - globalParameters["MergeFiles"] = True # F=store every solution and kernel in separate file; T=store all solutions in single file - - globalParameters["MaxFileName"] = 64 # If a file name would be longer than this, shorten it with a hash. -+globalParameters["SupportedISA"] = [(8,0,3), (9,0,0), (9,0,6), (9,0,8), (9,0,10), (10,1,0), (10,1,1), (10,1,2), (10,3,0),(10,3,1)] # assembly kernels writer supports these architectures --globalParameters["SupportedISA"] = [(8,0,3), (9,0,0), (9,0,6), (9,0,8), (9,0,10), (10,1,0), (10,1,1), (10,1,2), (10,3,0)] # assembly kernels writer supports these architectures - - globalParameters["GenerateManifestAndExit"] = False # Output manifest file with list of expected library objects and exit - globalParameters["ClientBuildPath"] = "0_Build" # subdirectory for host code build directory -@@ -265,7 +265,7 @@ - 'gfx906':'vega20', 'gfx906:xnack+':'vega20', 'gfx906:xnack-':'vega20', - 'gfx908':'arcturus','gfx908:xnack+':'arcturus', 'gfx908:xnack-':'arcturus', - 'gfx90a':'aldebaran', 'gfx90a:xnack+':'aldebaran', 'gfx90a:xnack-':'aldebaran', -+ 'gfx1010':'navi10', 'gfx1011':'navi11', 'gfx1012':'navi12', 'gfx1030':'navi21', 'gfx1031':'navi22', -- 'gfx1010':'navi10', 'gfx1011':'navi11', 'gfx1012':'navi12', 'gfx1030':'navi21' - } - - def getArchitectureName(gfxName): ---- Tensile-rocm-4.3.0/Tensile/KernelWriterAssembly.py 2021-04-22 22:35:33.000000000 +0800 -+++ Tensile-rocm-4.3.0/Tensile/KernelWriterAssembly.py 2021-09-30 17:27:06.696711144 +0800 -@@ -2257,7 +2257,7 @@ - - # integer i8x4 - elif kernel["ProblemType"]["DataType"].isInt8x4(): -+ if self.version == (9,0,6) or self.version == (9,0,8) or self.version == (9,0,10) or self.version == (10,3,0) or self.version == (10,3,1): -- if self.version == (9,0,6) or self.version == (9,0,8) or self.version == (9,0,10) or self.version == (10,3,0): - for b in range(0, kernel["ThreadTile1"]): - for a in range(0, kernel["ThreadTile0"]): - for iui in range(0, innerUnroll): ---- Tensile-rocm-4.3.0/Tensile/Source/CMakeLists.txt 2021-04-22 22:35:33.000000000 +0800 -+++ Tensile-rocm-4.3.0/Tensile/Source/CMakeLists.txt 2021-09-30 17:29:12.846345541 +0800 -@@ -51,7 +51,7 @@ - endif() - - if(CMAKE_CXX_COMPILER STREQUAL "hipcc") -+ set(TENSILE_GPU_ARCHS gfx803 gfx900 gfx906:xnack- gfx908:xnack- gfx90a:xnack- gfx1010 gfx1011 gfx1012 gfx1030 gfx1031 CACHE STRING "GPU architectures") -- set(TENSILE_GPU_ARCHS gfx803 gfx900 gfx906:xnack- gfx908:xnack- gfx90a:xnack- gfx1010 gfx1011 gfx1012 gfx1030 CACHE STRING "GPU architectures") - else() - set(TENSILE_GPU_ARCHS gfx803 gfx900 gfx906 gfx908 gfx90a CACHE STRING "GPU architectures") - endif() -@@ -104,7 +104,7 @@ - set_property( CACHE Tensile_CODE_OBJECT_VERSION PROPERTY STRINGS V2 V3) - set_property( CACHE Tensile_COMPILER PROPERTY STRINGS hipcc) - set(Tensile_ARCHITECTURE all CACHE STRING "Which GPU architecture to use") -+ set_property( CACHE Tensile_ARCHITECTURE PROPERTY STRINGS all gfx000 gfx900 gfx906:xnack- gfx908:xnack- gfx90a:xnack- gfx1010 gfx1011 gfx1012 gfx1030 gfx1031) -- set_property( CACHE Tensile_ARCHITECTURE PROPERTY STRINGS all gfx000 gfx900 gfx906:xnack- gfx908:xnack- gfx90a:xnack- gfx1010 gfx1011 gfx1012 gfx1030) - set_property( CACHE Tensile_LIBRARY_FORMAT PROPERTY STRINGS yaml msgpack) - - message( STATUS "Tensile_CODE_OBJECT_VERSION from Tensile/Source/CMakeLists.txt: ${Tensile_CODE_OBJECT_VERSION}") -@@ -206,7 +206,7 @@ - ${Tensile_RUNTIME_LANGUAGE} # HIP or OCL - ${Tensile_COMPILER} # hipcc - ${Tensile_CODE_OBJECT_VERSION} # V2 or V3 -+ ${Tensile_ARCHITECTURE} # all, gfx000, gfx803, gfx900, gfx906:xnack-, gfx908:xnack-, gfx90a:xnack-, gfx1010, gfx1011, gfx1012, gfx1030, gfx1031 -- ${Tensile_ARCHITECTURE} # all, gfx000, gfx803, gfx900, gfx906:xnack-, gfx908:xnack-, gfx90a:xnack-, gfx1010, gfx1011, gfx1012, gfx1030 - ${Tensile_LIBRARY_FORMAT} # yaml or msgpack - ${Tensile_MERGE_FILES} # ON or OFF - ${Tensile_SHORT_FILE_NAMES} # ON or OFF ---- Tensile-rocm-4.3.0/Tensile/Source/lib/include/Tensile/AMDGPU.hpp 2021-04-22 22:35:33.000000000 +0800 -+++ Tensile-rocm-4.3.0/Tensile/Source/lib/include/Tensile/AMDGPU.hpp 2021-09-30 17:48:11.277066515 +0800 -@@ -62,7 +62,8 @@ - gfx1010 = 1010, - gfx1011 = 1011, - gfx1012 = 1012, -+ gfx1030 = 1030, -+ gfx1031 = 1031 -- gfx1030 = 1030 - }; - - AMDGPU(); ---- Tensile-rocm-4.3.0/Tensile/Source/lib/include/Tensile/Serialization/Predicates.hpp 2021-04-22 22:35:33.000000000 +0800 -+++ Tensile-rocm-4.3.0/Tensile/Source/lib/include/Tensile/Serialization/Predicates.hpp 2021-09-30 17:31:26.075845249 +0800 -@@ -215,6 +215,7 @@ - iot::enumCase(io, value, "gfx1011", AMDGPU::Processor::gfx1011); - iot::enumCase(io, value, "gfx1012", AMDGPU::Processor::gfx1012); - iot::enumCase(io, value, "gfx1030", AMDGPU::Processor::gfx1030); -+ iot::enumCase(io, value, "gfx1031", AMDGPU::Processor::gfx1031); - } - }; - } // namespace Serialization ---- Tensile-rocm-4.3.0/Tensile/Source/lib/source/AMDGPU.cpp 2021-04-22 22:35:33.000000000 +0800 -+++ Tensile-rocm-4.3.0/Tensile/Source/lib/source/AMDGPU.cpp 2021-09-30 17:31:45.195486318 +0800 -@@ -82,6 +82,8 @@ - return stream << "gfx1012"; - case AMDGPU::Processor::gfx1030: - return stream << "gfx1030"; -+ case AMDGPU::Processor::gfx1031: -+ return stream << "gfx1031"; - } - return stream; - } ---- Tensile-rocm-4.3.0/Tensile/Source/TensileCreateLibrary.cmake 2021-04-22 22:35:33.000000000 +0800 -+++ Tensile-rocm-4.3.0/Tensile/Source/TensileCreateLibrary.cmake 2021-09-30 17:29:24.766121912 +0800 -@@ -138,7 +138,7 @@ - add_library(Tensile ${options} ${Tensile_SOURCE_FILES}) - # specify gpu targets - if( Tensile_ARCHITECTURE MATCHES "all" ) -+ set( Tensile_HIP_ISA "gfx803" "gfx900" "gfx906" "gfx908" "gfx1010" "gfx1011" "gfx1012" "gfx1030", "gfx1031") -- set( Tensile_HIP_ISA "gfx803" "gfx900" "gfx906" "gfx908" "gfx1010" "gfx1011" "gfx1012" "gfx1030") - else() - set( Tensile_HIP_ISA ${Tensile_ARCHITECTURE}) - endif() diff --git a/dev-util/Tensile/files/Tensile-4.3.0-hsaco-compile-specified-arch.patch b/dev-util/Tensile/files/Tensile-4.3.0-hsaco-compile-specified-arch.patch deleted file mode 100644 index 8e6753781a2a..000000000000 --- a/dev-util/Tensile/files/Tensile-4.3.0-hsaco-compile-specified-arch.patch +++ /dev/null @@ -1,96 +0,0 @@ -https://github.com/ROCmSoftwarePlatform/Tensile/issues/1395 -https://github.com/ROCmSoftwarePlatform/Tensile/pull/1398 - ---- a/Tensile/TensileCreateLibrary.py -+++ b/Tensile/TensileCreateLibrary.py -@@ -136,6 +136,35 @@ def which(p): - return candidate - return None - -+def splitArchs(): -+ # Helper for architecture -+ def isSupported(arch): -+ return globalParameters["AsmCaps"][arch]["SupportedISA"] and \ -+ globalParameters["AsmCaps"][arch]["SupportedSource"] -+ -+ if ";" in globalParameters["Architecture"]: -+ wantedArchs = globalParameters["Architecture"].split(";") -+ else: -+ wantedArchs = globalParameters["Architecture"].split("_") -+ archs = [] -+ cmdlineArchs = [] -+ if "all" in wantedArchs: -+ for arch in globalParameters['SupportedISA']: -+ if isSupported(arch): -+ if (arch == (9,0,6) or arch == (9,0,8) or arch == (9,0,10)): -+ if (arch == (9,0,10)): -+ archs += [gfxName(arch) + '-xnack+'] -+ cmdlineArchs += [gfxName(arch) + ':xnack+'] -+ archs += [gfxName(arch) + '-xnack-'] -+ cmdlineArchs += [gfxName(arch) + ':xnack-'] -+ else: -+ archs += [gfxName(arch)] -+ cmdlineArchs += [gfxName(arch)] -+ else: -+ for arch in wantedArchs: -+ archs += [re.sub(":", "-", arch)] -+ cmdlineArchs += [arch] -+ return archs, cmdlineArchs - - def buildSourceCodeObjectFile(CxxCompiler, outputPath, kernelFile): - buildPath = ensurePath(os.path.join(globalParameters['WorkingPath'], 'code_object_tmp')) -@@ -149,24 +178,8 @@ def buildSourceCodeObjectFile(CxxCompiler, outputPath, kernelFile): - objectFilename = base + '.o' - soFilename = base + '.so' - -- def isSupported(arch): -- return globalParameters["AsmCaps"][arch]["SupportedISA"] and \ -- globalParameters["AsmCaps"][arch]["SupportedSource"] -- - if (CxxCompiler == "hipcc"): -- archs = [] -- cmdlineArchs = [] -- for arch in globalParameters['SupportedISA']: -- if isSupported(arch): -- if (arch == (9,0,6) or arch == (9,0,8) or arch == (9,0,10)): -- if (arch == (9,0,10)): -- archs += [gfxName(arch) + '-xnack+'] -- cmdlineArchs += [gfxName(arch) + ':xnack+'] -- archs += [gfxName(arch) + '-xnack-'] -- cmdlineArchs += [gfxName(arch) + ':xnack-'] -- else: -- archs += [gfxName(arch)] -- cmdlineArchs += [gfxName(arch)] -+ archs, cmdlineArchs = splitArchs() - - archFlags = ['--offload-arch=' + arch for arch in cmdlineArchs] - -@@ -1063,11 +1076,6 @@ def buildObjectFileNames(solutionWriter, kernelWriterSource, kernelWriterAssembl - sourceKernels = list([k for k in kernels if k['KernelLanguage'] == 'Source']) - asmKernels = list([k for k in kernels if k['KernelLanguage'] == 'Assembly']) - -- # Helper for architecture -- def isSupported(arch): -- return globalParameters["AsmCaps"][arch]["SupportedISA"] and \ -- globalParameters["AsmCaps"][arch]["SupportedSource"] -- - # Build a list of kernel object names. - for kernel in sourceKernels: - sourceKernelNames += [kernelWriterSource.getKernelFileBase(kernel)] -@@ -1081,15 +1089,7 @@ def buildObjectFileNames(solutionWriter, kernelWriterSource, kernelWriterAssembl - - # Source based kernels are built for all supported architectures - if (cxxCompiler == 'hipcc'): -- sourceArchs = [] -- for arch in globalParameters['SupportedISA']: -- if isSupported(arch): -- if (arch == (9,0,6) or arch == (9,0,8) or arch == (9,0,10)): -- if (arch == (9,0,10)): -- sourceArchs += [gfxName(arch) + '-xnack+'] -- sourceArchs += [gfxName(arch) + '-xnack-'] -- else: -- sourceArchs += [gfxName(arch)] -+ sourceArchs, _ = splitArchs() - else: - raise RuntimeError("Unknown compiler %s" % cxxCompiler) - diff --git a/dev-util/Tensile/files/Tensile-4.3.0-use-ninja.patch b/dev-util/Tensile/files/Tensile-4.3.0-use-ninja.patch deleted file mode 100644 index 6645efd72fb1..000000000000 --- a/dev-util/Tensile/files/Tensile-4.3.0-use-ninja.patch +++ /dev/null @@ -1,21 +0,0 @@ -Index: Tensile-rocm-4.3.0/Tensile/ClientExecutable.py -=================================================================== ---- Tensile-rocm-4.3.0.orig/Tensile/ClientExecutable.py -+++ Tensile-rocm-4.3.0/Tensile/ClientExecutable.py -@@ -36,6 +36,7 @@ class CMakeEnvironment: - - args = ['cmake'] - args += itertools.chain.from_iterable([ ['-D', '{}={}'.format(key, value)] for key,value in self.options.items()]) -+ args += itertools.chain.from_iterable([ ['-G', 'Ninja'], [ '-D', 'CMAKE_EXPORT_COMPILE_COMMANDS=ON' ] ]) - args += [self.sourceDir] - - Common.print2(' '.join(args)) -@@ -43,7 +44,7 @@ class CMakeEnvironment: - subprocess.check_call(args, cwd=Common.ensurePath(self.buildDir)) - - def build(self): -- args = ['make', '-j'] -+ args = ['ninja'] - Common.print2(' '.join(args)) - with Common.ClientExecutionLock(): - subprocess.check_call(args, cwd=self.buildDir) diff --git a/dev-util/Tensile/files/Tensile-5.0.2-adjust-parallel.patch b/dev-util/Tensile/files/Tensile-5.0.2-adjust-parallel.patch deleted file mode 100644 index b96429232c78..000000000000 --- a/dev-util/Tensile/files/Tensile-5.0.2-adjust-parallel.patch +++ /dev/null @@ -1,65 +0,0 @@ -From 25b1621549f9b120462988913e657684645be79d Mon Sep 17 00:00:00 2001 -From: Lee Ringham <lringham@amd.com> -Date: Tue, 2 Nov 2021 14:43:07 -0600 -Subject: [PATCH] Added better control of parallelism to control memory usage - ---- - Tensile/Source/TensileCreateLibrary.cmake | 7 +++++-- - Tensile/cmake/TensileConfig.cmake | 5 +++++ - 2 files changed, 10 insertions(+), 2 deletions(-) - -diff --git a/Tensile/Source/TensileCreateLibrary.cmake b/Tensile/Source/TensileCreateLibrary.cmake -index 1dbacf41df..996f365bcf 100644 ---- a/Tensile/Source/TensileCreateLibrary.cmake -+++ b/Tensile/Source/TensileCreateLibrary.cmake -@@ -33,7 +33,8 @@ function(TensileCreateLibraryCmake - Tensile_LIBRARY_FORMAT - Tensile_MERGE_FILES - Tensile_SHORT_FILE_NAMES -- Tensile_LIBRARY_PRINT_DEBUG ) -+ Tensile_LIBRARY_PRINT_DEBUG -+ Tensile_CPU_THREADS ) - - # make Tensile_PACKAGE_LIBRARY and optional parameter - # to avoid breaking applications which us this -@@ -51,6 +52,7 @@ function(TensileCreateLibraryCmake - message(STATUS "Tensile_COMPILER from TensileCreateLibraryCmake : ${Tensile_COMPILER}") - message(STATUS "Tensile_ARCHITECTURE from TensileCreateLibraryCmake : ${Tensile_ARCHITECTURE}") - message(STATUS "Tensile_LIBRARY_FORMAT from TensileCreateLibraryCmake : ${Tensile_LIBRARY_FORMAT}") -+ message(STATUS "Tensile_CPU_THREADS from TensileCreateLibraryCmake : ${Tensile_CPU_THREADS}") - - #execute_process(COMMAND chmod 755 ${Tensile_ROOT}/bin/TensileCreateLibrary) - #execute_process(COMMAND chmod 755 ${Tensile_ROOT}/bin/Tensile) -@@ -89,7 +91,8 @@ function(TensileCreateLibraryCmake - set(Tensile_CREATE_COMMAND ${Tensile_CREATE_COMMAND} "--code-object-version=${Tensile_CODE_OBJECT_VERSION}") - set(Tensile_CREATE_COMMAND ${Tensile_CREATE_COMMAND} "--cxx-compiler=${Tensile_COMPILER}") - set(Tensile_CREATE_COMMAND ${Tensile_CREATE_COMMAND} "--library-format=${Tensile_LIBRARY_FORMAT}") -- -+ set(Tensile_CREATE_COMMAND ${Tensile_CREATE_COMMAND} "--jobs=${Tensile_CPU_THREADS}") -+ - # TensileLibraryWriter positional arguments - set(Tensile_CREATE_COMMAND ${Tensile_CREATE_COMMAND} - ${Tensile_LOGIC_PATH} -diff --git a/Tensile/cmake/TensileConfig.cmake b/Tensile/cmake/TensileConfig.cmake -index 0ac8517c2c..fa8cb96305 100644 ---- a/Tensile/cmake/TensileConfig.cmake -+++ b/Tensile/cmake/TensileConfig.cmake -@@ -112,6 +112,7 @@ function(TensileCreateLibraryFiles - LIBRARY_FORMAT - TENSILE_ROOT - VAR_PREFIX -+ CPU_THREADS - ) - - # Multi value settings -@@ -184,6 +185,10 @@ function(TensileCreateLibraryFiles - set(Options ${Options} "--cmake-cxx-compiler=${Tensile_COMPILER_PATH}") - endif() - -+ if(Tensile_CPU_THREADS) -+ set(Options ${Options} "--jobs=${Tensile_CPU_THREADS}") -+ endif() -+ - if(Tensile_LIBRARY_FORMAT) - set(Options ${Options} "--library-format=${Tensile_LIBRARY_FORMAT}") - if(Tensile_LIBRARY_FORMAT MATCHES "yaml") diff --git a/dev-util/Tensile/files/Tensile-5.0.2-gentoopath.patch b/dev-util/Tensile/files/Tensile-5.1.3-gentoopath.patch index b8b59af5184e..2f41bf3aa715 100644 --- a/dev-util/Tensile/files/Tensile-5.0.2-gentoopath.patch +++ b/dev-util/Tensile/files/Tensile-5.1.3-gentoopath.patch @@ -17,7 +17,7 @@ Index: Tensile-rocm-5.0.2/Tensile/Source/lib/CMakeLists.txt if(TENSILE_USE_LLVM) - find_package(LLVM 13.0 QUIET CONFIG) -+ find_package(LLVM PATHS @GENTOO_PORTAGE_EPREFIX@/usr/lib/llvm/roc QUIET CONFIG) ++ find_package(LLVM PATHS @LLVM_PATH@ QUIET CONFIG) if(NOT LLVM_FOUND) find_package(LLVM 12.0 QUIET CONFIG) if(NOT LLVM_FOUND) @@ -69,7 +69,7 @@ Index: Tensile-rocm-5.0.2/HostLibraryTests/CMakeLists.txt if(TENSILE_USE_LLVM) - find_package(LLVM 13.0 QUIET CONFIG) -+ find_package(LLVM PATHS @GENTOO_PORTAGE_EPREFIX@/usr/lib/llvm/roc QUIET CONFIG) ++ find_package(LLVM PATHS @LLVM_PATH@ QUIET CONFIG) if(NOT LLVM_FOUND) find_package(LLVM 12.0 QUIET CONFIG) if(NOT LLVM_FOUND) diff --git a/dev-util/Tensile/files/Tensile-5.4.2-fix-arch-parse.patch b/dev-util/Tensile/files/Tensile-5.4.2-fix-arch-parse.patch new file mode 100644 index 000000000000..14263e89ef84 --- /dev/null +++ b/dev-util/Tensile/files/Tensile-5.4.2-fix-arch-parse.patch @@ -0,0 +1,12 @@ +Index: Tensile-rocm-5.3.3/Tensile/Tensile.py +=================================================================== +--- Tensile-rocm-5.3.3.orig/Tensile/Tensile.py ++++ Tensile-rocm-5.3.3/Tensile/Tensile.py +@@ -101,7 +101,6 @@ def addCommonArguments(argParser): + Allows the --global-parameters option to specify any parameters from the command line. + """ + (key, value) = par.split("=") +- value = eval(value) + return (key, value) + + argParser.add_argument("-d", "--device", dest="device", type=int, \ diff --git a/dev-util/Tensile/files/Tensile-5.4.2-gentoopath.patch b/dev-util/Tensile/files/Tensile-5.4.2-gentoopath.patch new file mode 100644 index 000000000000..eda9b0307c52 --- /dev/null +++ b/dev-util/Tensile/files/Tensile-5.4.2-gentoopath.patch @@ -0,0 +1,67 @@ +Fix incorrect path in scripts and CMakeLists, in order to build tensile_client +=================================================================== +Index: Tensile-rocm-5.4.2/Tensile/Source/lib/CMakeLists.txt +=================================================================== +--- Tensile-rocm-5.4.2.orig/Tensile/Source/lib/CMakeLists.txt ++++ Tensile-rocm-5.4.2/Tensile/Source/lib/CMakeLists.txt +@@ -51,7 +51,7 @@ set(tensile_sources ${tensile_sources} + ) + + if(TENSILE_USE_LLVM) +- find_package(LLVM 13.0 QUIET CONFIG) ++ find_package(LLVM PATHS @LLVM_PATH@ QUIET CONFIG) + if(NOT LLVM_FOUND) + find_package(LLVM 12.0 QUIET CONFIG) + if(NOT LLVM_FOUND) +Index: Tensile-rocm-5.4.2/Tensile/Common.py +=================================================================== +--- Tensile-rocm-5.4.2.orig/Tensile/Common.py ++++ Tensile-rocm-5.4.2/Tensile/Common.py +@@ -1950,7 +1950,7 @@ def assignGlobalParameters( config ): + else: + print2(" %24s: %8s (unspecified)" % (key, defaultValue)) + +- globalParameters["ROCmPath"] = "/opt/rocm" ++ globalParameters["ROCmPath"] = "@GENTOO_PORTAGE_EPREFIX@/usr" + if "ROCM_PATH" in os.environ: + globalParameters["ROCmPath"] = os.environ.get("ROCM_PATH") + if "TENSILE_ROCM_PATH" in os.environ: +Index: Tensile-rocm-5.4.2/Tensile/ClientWriter.py +=================================================================== +--- Tensile-rocm-5.4.2.orig/Tensile/ClientWriter.py ++++ Tensile-rocm-5.4.2/Tensile/ClientWriter.py +@@ -180,7 +180,7 @@ def runClient(libraryLogicPath, forBench + def getBuildClientLibraryScript(buildPath, libraryLogicPath): + callCreateLibraryCmd = ["python"] if os.name == "nt" else [] + +- callCreateLibraryCmd += [os.path.join(globalParameters["ScriptPath"] , "bin", "TensileCreateLibrary")] ++ callCreateLibraryCmd += ["TensileCreateLibrary"] + + if globalParameters["MergeFiles"]: + callCreateLibraryCmd += ["--merge-files"] +Index: Tensile-rocm-5.4.2/Tensile/GenerateSummations.py +=================================================================== +--- Tensile-rocm-5.4.2.orig/Tensile/GenerateSummations.py ++++ Tensile-rocm-5.4.2/Tensile/GenerateSummations.py +@@ -59,7 +59,7 @@ def createLibraryForBenchmark(logicPath, + Selection. + """ + +- pythonExePath = os.path.join(os.path.dirname(os.path.realpath(__file__)), "bin", "TensileCreateLibrary") ++ pythonExePath = "TensileCreateLibrary" + args = [pythonExePath, \ + "--merge-files", "--new-client-only", "--no-short-file-names", "--no-library-print-debug", \ + "--architecture=all", "--code-object-version=V3", "--cxx-compiler=hipcc", "--library-format=yaml", \ +Index: Tensile-rocm-5.4.2/HostLibraryTests/CMakeLists.txt +=================================================================== +--- Tensile-rocm-5.4.2.orig/HostLibraryTests/CMakeLists.txt ++++ Tensile-rocm-5.4.2/HostLibraryTests/CMakeLists.txt +@@ -139,7 +139,7 @@ set(test_sources ${test_sources} + ) + + if(TENSILE_USE_LLVM) +- find_package(LLVM 13.0 QUIET CONFIG) ++ find_package(LLVM PATHS @LLVM_PATH@ QUIET CONFIG) + if(NOT LLVM_FOUND) + find_package(LLVM 12.0 QUIET CONFIG) + if(NOT LLVM_FOUND) diff --git a/dev-util/Tensile/files/Tensile-5.4.2-gfx1031.patch b/dev-util/Tensile/files/Tensile-5.4.2-gfx1031.patch new file mode 100644 index 000000000000..1f703c07c3ec --- /dev/null +++ b/dev-util/Tensile/files/Tensile-5.4.2-gfx1031.patch @@ -0,0 +1,88 @@ +Index: Tensile-rocm-5.3.3/pytest.ini +=================================================================== +--- Tensile-rocm-5.3.3.orig/pytest.ini ++++ Tensile-rocm-5.3.3/pytest.ini +@@ -88,6 +88,7 @@ markers = + xfail-gfx1011: architecture + xfail-gfx1012: architecture + xfail-gfx1030: architecture ++ xfail-gfx1031: architecture + xfail-gfx1100: architecture + xfail-gfx1101: architecture + xfail-gfx1102: architecture +@@ -100,6 +101,7 @@ markers = + skip-gfx1011: architecture + skip-gfx1012: architecture + skip-gfx1030: architecture ++ skip-gfx1031: architecture + skip-gfx1100: architecture + skip-gfx1101: architecture + skip-gfx1102: architecture +Index: Tensile-rocm-5.3.3/Tensile/Common.py +=================================================================== +--- Tensile-rocm-5.3.3.orig/Tensile/Common.py ++++ Tensile-rocm-5.3.3/Tensile/Common.py +@@ -201,7 +201,7 @@ globalParameters["MergeFiles"] = True + globalParameters["NumMergedFiles"] = 1 # The number of files that kernels should be split between when merging + + globalParameters["MaxFileName"] = 64 # If a file name would be longer than this, shorten it with a hash. +-globalParameters["SupportedISA"] = [(8,0,3), (9,0,0), (9,0,6), (9,0,8), (9,0,10), (10,1,0), (10,1,1), (10,1,2), (10,3,0), (11,0,0), (11,0,1), (11,0,2)] # assembly kernels writer supports these architectures ++globalParameters["SupportedISA"] = [(8,0,3), (9,0,0), (9,0,6), (9,0,8), (9,0,10), (10,1,0), (10,1,1), (10,1,2), (10,3,0), (10,3,1), (11,0,0), (11,0,1), (11,0,2)] # assembly kernels writer supports these architectures + + globalParameters["GenerateManifestAndExit"] = False # Output manifest file with list of expected library objects and exit + globalParameters["NewClient"] = 2 # Old client deprecated: NewClient must be set to 2. +@@ -271,7 +271,7 @@ architectureMap = { + 'gfx906':'vega20', 'gfx906:xnack+':'vega20', 'gfx906:xnack-':'vega20', + 'gfx908':'arcturus','gfx908:xnack+':'arcturus', 'gfx908:xnack-':'arcturus', + 'gfx90a':'aldebaran', 'gfx90a:xnack+':'aldebaran', 'gfx90a:xnack-':'aldebaran', +- 'gfx1010':'navi10', 'gfx1011':'navi12', 'gfx1012':'navi14', 'gfx1030':'navi21', ++ 'gfx1010':'navi10', 'gfx1011':'navi12', 'gfx1012':'navi14', 'gfx1030':'navi21', 'gfx1031':'navi22', + 'gfx1100':'navi31', 'gfx1101':'navi32', 'gfx1102':'navi33' + } + +Index: Tensile-rocm-5.3.3/Tensile/Source/CMakeLists.txt +=================================================================== +--- Tensile-rocm-5.3.3.orig/Tensile/Source/CMakeLists.txt ++++ Tensile-rocm-5.3.3/Tensile/Source/CMakeLists.txt +@@ -51,7 +51,7 @@ if(NOT DEFINED CXX_VERSION_STRING) + endif() + + if(CMAKE_CXX_COMPILER STREQUAL "hipcc") +- set(TENSILE_GPU_ARCHS gfx803 gfx900 gfx906:xnack- gfx908:xnack- gfx90a:xnack- gfx1010 gfx1011 gfx1012 gfx1030 gfx1100 gfx1101 gfx1102 CACHE STRING "GPU architectures") ++ set(TENSILE_GPU_ARCHS gfx803 gfx900 gfx906:xnack- gfx908:xnack- gfx90a:xnack- gfx1010 gfx1011 gfx1012 gfx1030 gfx1031 gfx1100 gfx1101 gfx1102 CACHE STRING "GPU architectures") + else() + set(TENSILE_GPU_ARCHS gfx803 gfx900 gfx906 gfx908 gfx90a gfx1010 gfx1011 gfx1012 gfx1030 gfx1100 gfx1101 gfx1102 CACHE STRING "GPU architectures") + endif() +Index: Tensile-rocm-5.3.3/Tensile/Source/lib/include/Tensile/AMDGPU.hpp +=================================================================== +--- Tensile-rocm-5.3.3.orig/Tensile/Source/lib/include/Tensile/AMDGPU.hpp ++++ Tensile-rocm-5.3.3/Tensile/Source/lib/include/Tensile/AMDGPU.hpp +@@ -63,6 +63,7 @@ namespace Tensile + gfx1011 = 1011, + gfx1012 = 1012, + gfx1030 = 1030, ++ gfx1031 = 1031, + gfx1100 = 1100, + gfx1101 = 1101, + gfx1102 = 1102 +@@ -90,6 +91,8 @@ namespace Tensile + return "gfx1012"; + case AMDGPU::Processor::gfx1030: + return "gfx1030"; ++ case AMDGPU::Processor::gfx1031: ++ return "gfx1032"; + case AMDGPU::Processor::gfx1100: + return "gfx1100"; + case AMDGPU::Processor::gfx1101: +Index: Tensile-rocm-5.3.3/Tensile/Source/lib/include/Tensile/Serialization/Predicates.hpp +=================================================================== +--- Tensile-rocm-5.3.3.orig/Tensile/Source/lib/include/Tensile/Serialization/Predicates.hpp ++++ Tensile-rocm-5.3.3/Tensile/Source/lib/include/Tensile/Serialization/Predicates.hpp +@@ -215,6 +215,7 @@ namespace Tensile + iot::enumCase(io, value, "gfx1011", AMDGPU::Processor::gfx1011); + iot::enumCase(io, value, "gfx1012", AMDGPU::Processor::gfx1012); + iot::enumCase(io, value, "gfx1030", AMDGPU::Processor::gfx1030); ++ iot::enumCase(io, value, "gfx1031", AMDGPU::Processor::gfx1031); + iot::enumCase(io, value, "gfx1100", AMDGPU::Processor::gfx1100); + iot::enumCase(io, value, "gfx1101", AMDGPU::Processor::gfx1101); + iot::enumCase(io, value, "gfx1102", AMDGPU::Processor::gfx1102); diff --git a/dev-util/Tensile/files/Tensile-5.4.2-use-ninja.patch b/dev-util/Tensile/files/Tensile-5.4.2-use-ninja.patch new file mode 100644 index 000000000000..f88bf00665cb --- /dev/null +++ b/dev-util/Tensile/files/Tensile-5.4.2-use-ninja.patch @@ -0,0 +1,21 @@ +Index: Tensile-rocm-5.4.2/Tensile/ClientExecutable.py +=================================================================== +--- Tensile-rocm-5.4.2.orig/Tensile/ClientExecutable.py ++++ Tensile-rocm-5.4.2/Tensile/ClientExecutable.py +@@ -44,6 +44,7 @@ class CMakeEnvironment: + args = ['cmake'] + args += ['-G', 'Ninja'] if (os.name == 'nt') else [] + args += itertools.chain.from_iterable([ ['-D{}={}'.format(key, value)] for key,value in self.options.items()]) ++ args += itertools.chain.from_iterable([ ['-G', 'Ninja'], [ '-D', 'CMAKE_EXPORT_COMPILE_COMMANDS=ON' ] ]) + args += [self.sourceDir] + args = [cmake_path(arg) for arg in args] + +@@ -53,7 +54,7 @@ class CMakeEnvironment: + subprocess.check_output(args, stderr=subprocess.STDOUT, cwd=Common.ensurePath(self.buildDir)) + + def build(self): +- args = [('ninja' if (os.name == "nt") else 'make'), f'-j{CPUThreadCount()}'] ++ args = [('ninja'), f'-j{CPUThreadCount()}'] + Common.print2(' '.join(args)) + with Common.ClientExecutionLock(): + # change to use check_output to force windows cmd block util command finish diff --git a/dev-util/Tensile/files/Tensile-5.7.1-expand-isa-compatibility.patch b/dev-util/Tensile/files/Tensile-5.7.1-expand-isa-compatibility.patch new file mode 100644 index 000000000000..ebfc537c5c4d --- /dev/null +++ b/dev-util/Tensile/files/Tensile-5.7.1-expand-isa-compatibility.patch @@ -0,0 +1,49 @@ +Allow Tensile to load the compatible kernels when running on +architectures compatible with those ISAs. + +Based on patch from Cordell Bloor <cgmb@slerp.xyz> +https://salsa.debian.org/rocm-team/rocblas/-/blob/master/debian/patches/0012-expand-isa-compatibility.patch + +--- a/Tensile/Source/lib/source/hip/HipHardware.cpp ++++ b/Tensile/Source/lib/source/hip/HipHardware.cpp +@@ -24,6 +24,8 @@ + * + *******************************************************************************/ + ++#include <cstring> ++ + #include <Tensile/AMDGPU.hpp> + #include <Tensile/hip/HipHardware.hpp> + #include <Tensile/hip/HipUtils.hpp> +@@ -56,6 +58,31 @@ namespace Tensile + { + hipDeviceProp_t prop; + HIP_CHECK_EXC(hipGetDeviceProperties(&prop, deviceId)); ++ // coerce to compatible arch ++ switch(prop.gcnArch) ++ { ++ case 902: ++ case 909: ++ case 912: ++ prop.gcnArch = 900; ++ std::strcpy(prop.gcnArchName, "gfx900"); ++ break; ++ case 1011: ++ case 1012: ++ case 1013: ++ prop.gcnArch = 1010; ++ std::strcpy(prop.gcnArchName, "gfx1010"); ++ break; ++ case 1031: ++ case 1032: ++ case 1033: ++ case 1034: ++ case 1035: ++ case 1036: ++ prop.gcnArch = 1030; ++ std::strcpy(prop.gcnArchName, "gfx1030"); ++ break; ++ } + #if HIP_VERSION >= 50220730 + int hip_version; + HIP_CHECK_EXC(hipRuntimeGetVersion(&hip_version)); diff --git a/dev-util/Tensile/files/Tensile-5.7.1-fix-msgpack-dependency.patch b/dev-util/Tensile/files/Tensile-5.7.1-fix-msgpack-dependency.patch new file mode 100644 index 000000000000..f6f3f0e24c1a --- /dev/null +++ b/dev-util/Tensile/files/Tensile-5.7.1-fix-msgpack-dependency.patch @@ -0,0 +1,29 @@ +Gentoo uses "msgpack-cxx" as library name and adds important +MSGPACK_NO_BOOST compile definition when msgpack-cxx us installed with USE=-boost +--- a/Tensile/Source/lib/CMakeLists.txt ++++ b/Tensile/Source/lib/CMakeLists.txt +@@ -98,22 +98,9 @@ if(TENSILE_USE_LLVM OR TENSILE_USE_MSGPACK) + endif() + + if(TENSILE_USE_MSGPACK) +- find_package(msgpack REQUIRED) ++ find_package(msgpack-cxx REQUIRED) + target_compile_definitions(TensileHost PUBLIC -DTENSILE_MSGPACK=1) +- +- if(TARGET msgpackc-cxx) +- get_target_property(msgpack_inc msgpackc-cxx INTERFACE_INCLUDE_DIRECTORIES) +- elseif(TARGET msgpackc) +- get_target_property(msgpack_inc msgpackc INTERFACE_INCLUDE_DIRECTORIES) +- endif() +- +- if(DEFINED msgpack_inc) +- # include C++ headers manually +- # External header includes included as system files +- target_include_directories(TensileHost +- SYSTEM PRIVATE $<BUILD_INTERFACE:${msgpack_inc}> +- ) +- endif() ++ target_link_libraries(TensileHost PRIVATE msgpack-cxx) + endif() + + if(TENSILE_USE_LLVM) diff --git a/dev-util/Tensile/files/Tensile-5.7.1-gentoopath.patch b/dev-util/Tensile/files/Tensile-5.7.1-gentoopath.patch new file mode 100644 index 000000000000..c8bc360cc39b --- /dev/null +++ b/dev-util/Tensile/files/Tensile-5.7.1-gentoopath.patch @@ -0,0 +1,57 @@ +Fix incorrect path in scripts and CMakeLists, in order to build tensile_client + +--- a/HostLibraryTests/CMakeLists.txt ++++ b/HostLibraryTests/CMakeLists.txt +@@ -139,7 +139,7 @@ set(test_sources ${test_sources} + ) + + if(TENSILE_USE_LLVM) +- find_package(LLVM REQUIRED CONFIG) ++ find_package(LLVM PATHS @LLVM_PATH@ REQUIRED CONFIG) + + set(test_sources ${test_sources} + ContractionLibraryLoading_test.cpp +--- a/Tensile/ClientWriter.py ++++ b/Tensile/ClientWriter.py +@@ -176,7 +176,7 @@ def runClient(libraryLogicPath, forBenchmark, enableTileSelection, configPaths=N + def getBuildClientLibraryScript(buildPath, libraryLogicPath): + callCreateLibraryCmd = ["python"] if os.name == "nt" else [] + +- callCreateLibraryCmd += [os.path.join(globalParameters["ScriptPath"] , "bin", "TensileCreateLibrary")] ++ callCreateLibraryCmd += ["TensileCreateLibrary"] + + if globalParameters["MergeFiles"]: + callCreateLibraryCmd += ["--merge-files"] +--- a/Tensile/Common.py ++++ b/Tensile/Common.py +@@ -2108,7 +2108,7 @@ def assignGlobalParameters( config ): + else: + print2(" %24s: %8s (unspecified)" % (key, defaultValue)) + +- globalParameters["ROCmPath"] = "/opt/rocm" ++ globalParameters["ROCmPath"] = "@GENTOO_PORTAGE_EPREFIX@/usr" + if "ROCM_PATH" in os.environ: + globalParameters["ROCmPath"] = os.environ.get("ROCM_PATH") + if "TENSILE_ROCM_PATH" in os.environ: +--- a/Tensile/GenerateSummations.py ++++ b/Tensile/GenerateSummations.py +@@ -59,7 +59,7 @@ def createLibraryForBenchmark(logicPath, libraryPath, currentPath): + Selection. + """ + +- pythonExePath = os.path.join(os.path.dirname(os.path.realpath(__file__)), "bin", "TensileCreateLibrary") ++ pythonExePath = "TensileCreateLibrary" + args = [pythonExePath, \ + "--merge-files", "--new-client-only", "--no-short-file-names", "--no-library-print-debug", \ + "--architecture=all", "--cxx-compiler=hipcc", "--library-format=yaml", \ +--- a/Tensile/Source/lib/CMakeLists.txt ++++ b/Tensile/Source/lib/CMakeLists.txt +@@ -52,7 +52,7 @@ set(tensile_sources ${tensile_sources} + ) + + if(TENSILE_USE_LLVM) +- find_package(LLVM REQUIRED CONFIG) ++ find_package(LLVM PATHS @LLVM_PATH@ REQUIRED CONFIG) + + set(tensile_sources ${tensile_sources} + source/llvm/YAML.cpp diff --git a/dev-util/Tensile/files/Tensile-change-cmake-name-for-msgpack-cxx-6-release.patch b/dev-util/Tensile/files/Tensile-change-cmake-name-for-msgpack-cxx-6-release.patch new file mode 100644 index 000000000000..d8ac203781da --- /dev/null +++ b/dev-util/Tensile/files/Tensile-change-cmake-name-for-msgpack-cxx-6-release.patch @@ -0,0 +1,13 @@ +Yiyang's original patch for msgpack-cxx-5 adapted to version 6. + +--- a/Tensile/Source/lib/CMakeLists.txt ++++ b/Tensile/Source/lib/CMakeLists.txt +@@ -97,7 +97,7 @@ if(TENSILE_USE_LLVM OR TENSILE_USE_MSGPACK) + endif() + + if(TENSILE_USE_MSGPACK) +- find_package(msgpack REQUIRED) ++ find_package(msgpack-cxx REQUIRED) + target_compile_definitions(TensileHost PUBLIC -DTENSILE_MSGPACK=1) + + if(TARGET msgpackc-cxx) diff --git a/dev-util/Tensile/metadata.xml b/dev-util/Tensile/metadata.xml index 97a4d2e3c600..32364cb13b06 100644 --- a/dev-util/Tensile/metadata.xml +++ b/dev-util/Tensile/metadata.xml @@ -8,4 +8,7 @@ <upstream> <remote-id type="github">ROCmSoftwarePlatform/Tensile</remote-id> </upstream> + <use> + <flag name="client">Build and install tensile_client executable to run benchmarks and tune GPU GEMM</flag> + </use> </pkgmetadata> |