commit: 1a382d070b821d8247e0ea87e55faca0f97d3d10 Author: Paul Zander <negril.nx+gentoo <AT> gmail <DOT> com> AuthorDate: Thu Jan 30 22:08:31 2025 +0000 Commit: Sam James <sam <AT> gentoo <DOT> org> CommitDate: Fri Mar 7 13:21:43 2025 +0000 URL: https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=1a382d07
media-libs/oidn: add 2.3.2 Closes: https://github.com/gentoo/gentoo/pull/39854 Closes: https://github.com/gentoo/gentoo/pull/40320 Closes: https://bugs.gentoo.org/950261 Signed-off-by: Paul Zander <negril.nx+gentoo <AT> gmail.com> Closes: https://github.com/gentoo/gentoo/pull/40390 Signed-off-by: Sam James <sam <AT> gentoo.org> media-libs/oidn/Manifest | 2 +- .../files/oidn-2.3.1-composable-kernel-api.patch | 28 +++ .../oidn-2.3.1-system-composable-kernel.patch | 199 +++++++++++++++++++++ .../oidn/{oidn-2.3.1.ebuild => oidn-2.3.2.ebuild} | 41 +++-- 4 files changed, 254 insertions(+), 16 deletions(-) diff --git a/media-libs/oidn/Manifest b/media-libs/oidn/Manifest index f23995017019..f163792ea8c6 100644 --- a/media-libs/oidn/Manifest +++ b/media-libs/oidn/Manifest @@ -1,3 +1,3 @@ DIST oidn-2.2.2.tar.gz 49006605 BLAKE2B e3d32cc2112ded280e642c5d69fb2ef14db59df4f5224bff79e82ede207899c9dbfa43dc663ad9ccc15a6497e363df0b91b43660729962c410b7ec75ac1b3465 SHA512 eca18e9e22a40bc1381e75985e5f5e79f48cc6d168a2a703e54540967646487d347db290459e9a612d8ab4924e8d3c83d84664fd5318cca4f17745da1e47cf50 DIST oidn-2.3.0.tar.gz 75300547 BLAKE2B 88a5922b4db63c48ab69047dca8d1077205a6d7a2112971badb0cd35e0b770017f5cdcc95d2ef07016fe2e990e3e5806ac6a0800b67086b958948c542f0dff6f SHA512 4f68ce3b0da37e1861090d60d7cd823046f325c45a8e2c3f4098eb3b4f7ec676d2acbb9c138ff60a9f83188dd0dcbaba42106aad96beafc5e65045c77c00f2f8 -DIST oidn-2.3.1.tar.gz 75315393 BLAKE2B d031e89924f82514994f61909d949b5dc55585cca6fbe26aba2e587f2afd03a6ae195aa93c0bbe67554224bab97ab552a4f33723786068abb82c1f6dde58e2b4 SHA512 ed3c99ba3d6634d8ca55786e4e62fe78cf6941e80736e6fcdf8e4f7b08aa45d1a546692bc64fedead31598b32b050a7d8a7e2cd1a6f3c55d645eeb175e440042 +DIST oidn-2.3.2.tar.gz 75315672 BLAKE2B d271a296ee8f72493ad109e1e13dc8142f8e27ed878e08dc9baa7fc04e936ef8604bedf7fa65df278758f2c12c9e66cd517aab16b422b2b99d68aff6c7e56587 SHA512 bba24f712bdd9f35a4e2f60a0b86b1b48eac40bf40a898cd8c2a7571e52f5d0b0055fcd8b6647a9ff062e2680ded63b66aa916a5a4b428de114c0a218f7783ea diff --git a/media-libs/oidn/files/oidn-2.3.1-composable-kernel-api.patch b/media-libs/oidn/files/oidn-2.3.1-composable-kernel-api.patch new file mode 100644 index 000000000000..80f6bf5e9f26 --- /dev/null +++ b/media-libs/oidn/files/oidn-2.3.1-composable-kernel-api.patch @@ -0,0 +1,28 @@ +diff --git a/devices/hip/ck_conv_wmma.cpp b/devices/hip/ck_conv_wmma.cpp +index 144374d..1bb2f7f 100644 +--- a/devices/hip/ck_conv_wmma.cpp ++++ b/devices/hip/ck_conv_wmma.cpp +@@ -48,19 +48,20 @@ OIDN_NAMESPACE_BEGIN + OutLayout, // ELayout + InDataType, // ADataType + WeiDataType, // BDataType +- ck::Tuple<BiasDataType>, // DsDataType +- OutDataType, // EDataType + AccDataType, // AccDataType + CShuffleDataType, // CShuffleDataType ++ ck::Tuple<BiasDataType>, // DsDataType ++ OutDataType, // EDataType + InElementOp, // AElementwiseOperation + WeiElementOp, // BElementwiseOperation + OutElementOp, // CDEElementwiseOperation + ConvSpec, // ConvForwardSpecialization + GemmSpec, // GemmSpecialization ++ 1, // NumGemmKPrefetchStage + 256, // BlockSize + 128, // MPerBlock + 64, // NPerBlock +- 4, // K0PerBlock ++ 32, // KPerBlock + 8, // K1 + 16, // MPerWMMA + 16, // NPerWMMA diff --git a/media-libs/oidn/files/oidn-2.3.1-system-composable-kernel.patch b/media-libs/oidn/files/oidn-2.3.1-system-composable-kernel.patch new file mode 100644 index 000000000000..e1b96181ec91 --- /dev/null +++ b/media-libs/oidn/files/oidn-2.3.1-system-composable-kernel.patch @@ -0,0 +1,199 @@ +diff --git a/devices/hip/CMakeLists.txt b/devices/hip/CMakeLists.txt +index e9a9bec..b738d5b 100644 +--- a/devices/hip/CMakeLists.txt ++++ b/devices/hip/CMakeLists.txt +@@ -53,12 +53,6 @@ set_target_properties(OpenImageDenoise_device_hip PROPERTIES + CXX_STANDARD 17 + ) + +-target_include_directories(OpenImageDenoise_device_hip +- PRIVATE +- "${OIDN_ROOT_SOURCE_DIR}/external/composable_kernel/include" +- "${OIDN_ROOT_SOURCE_DIR}/external/composable_kernel/library/include" +-) +- + set_source_files_properties( + ck_conv_wmma.cpp + PROPERTIES COMPILE_FLAGS "-mcumode -mno-wavefrontsize64" # Navi3x optimizations +diff --git a/devices/hip/ck_conv_dl.cpp b/devices/hip/ck_conv_dl.cpp +index 1aacfe8..e99da53 100644 +--- a/devices/hip/ck_conv_dl.cpp ++++ b/devices/hip/ck_conv_dl.cpp +@@ -3,7 +3,7 @@ + + #include "hip_conv.h" + #include "ck_conv.h" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + OIDN_NAMESPACE_BEGIN + +diff --git a/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp16.cpp b/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp16.cpp +index db5a7f0..2f384cb 100644 +--- a/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp16.cpp ++++ b/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp16.cpp +@@ -3,7 +3,7 @@ + + #include "convnd_fwd_dl_common.hpp" + +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + #include "ck/library/utility/convolution_host_tensor_descriptor_helper.hpp" + +diff --git a/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp32.cpp b/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp32.cpp +index 964d784..aa56666 100644 +--- a/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp32.cpp ++++ b/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_fp32.cpp +@@ -3,7 +3,7 @@ + + #include "convnd_fwd_dl_common.hpp" + +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + #include "ck/library/utility/convolution_host_tensor_descriptor_helper.hpp" + +diff --git a/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_int8.cpp b/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_int8.cpp +index b0cd88f..15b554c 100644 +--- a/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_int8.cpp ++++ b/external/composable_kernel/example/09_convnd_fwd/convnd_fwd_dl_int8.cpp +@@ -3,7 +3,7 @@ + + #include "convnd_fwd_dl_common.hpp" + +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + #include "ck/library/utility/convolution_host_tensor_descriptor_helper.hpp" + +diff --git a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp +index 5c445d9..49d2ba3 100644 +--- a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp ++++ b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp +@@ -2,7 +2,7 @@ + // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. + + #include "common.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + using InDataType = int8_t; + using WeiDataType = int8_t; +diff --git a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp +index 0ff85f0..661d744 100644 +--- a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp ++++ b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp +@@ -2,7 +2,7 @@ + // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. + + #include "common.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + using InDataType = int8_t; + using WeiDataType = int8_t; +diff --git a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp +index f8f996d..65cded3 100644 +--- a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp ++++ b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp +@@ -2,7 +2,7 @@ + // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. + + #include "common.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + using InDataType = int8_t; + using WeiDataType = int8_t; +diff --git a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp +index 3b25fec..ca5f217 100644 +--- a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp ++++ b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp +@@ -2,7 +2,7 @@ + // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. + + #include "common.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + using InDataType = int8_t; + using WeiDataType = int8_t; +diff --git a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perchannel_quantization_int8.cpp b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perchannel_quantization_int8.cpp +index a98a1e2..88b9ccc 100644 +--- a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perchannel_quantization_int8.cpp ++++ b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perchannel_quantization_int8.cpp +@@ -2,7 +2,7 @@ + // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. + + #include "common.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + using InDataType = int8_t; + using WeiDataType = int8_t; +diff --git a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perlayer_quantization_int8.cpp b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perlayer_quantization_int8.cpp +index 262594d..f5678e7 100644 +--- a/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perlayer_quantization_int8.cpp ++++ b/external/composable_kernel/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perlayer_quantization_int8.cpp +@@ -2,7 +2,7 @@ + // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. + + #include "common.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + using InDataType = int8_t; + using WeiDataType = int8_t; +diff --git a/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f16_instance.cpp b/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f16_instance.cpp +index fc18b3c..020d93e 100644 +--- a/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f16_instance.cpp ++++ b/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f16_instance.cpp +@@ -5,7 +5,7 @@ + + #include "ck/ck.hpp" + #include "ck/tensor_operation/gpu/device/tensor_layout.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp" + #include "ck/tensor_operation/gpu/device/convolution_forward_specialization.hpp" + #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp" +diff --git a/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f32_instance.cpp b/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f32_instance.cpp +index 648b396..42825ce 100644 +--- a/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f32_instance.cpp ++++ b/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_f32_instance.cpp +@@ -5,7 +5,7 @@ + + #include "ck/ck.hpp" + #include "ck/tensor_operation/gpu/device/tensor_layout.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp" + #include "ck/tensor_operation/gpu/device/convolution_forward_specialization.hpp" + #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp" +diff --git a/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_int8_instance.cpp b/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_int8_instance.cpp +index 1cb5d06..ab74235 100644 +--- a/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_int8_instance.cpp ++++ b/external/composable_kernel/library/src/tensor_operation_instance/gpu/grouped_conv2d_fwd/device_grouped_conv2d_fwd_dl_gnhwc_gkyxc_gnhwk_int8_instance.cpp +@@ -5,7 +5,7 @@ + + #include "ck/ck.hpp" + #include "ck/tensor_operation/gpu/device/tensor_layout.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + #include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp" + #include "ck/tensor_operation/gpu/device/convolution_forward_specialization.hpp" + #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp" +diff --git a/external/composable_kernel/library/src/tensor_operation_instance/gpu/quantization/conv2d_fwd/device_conv2d_dl_int8_instance.hpp b/external/composable_kernel/library/src/tensor_operation_instance/gpu/quantization/conv2d_fwd/device_conv2d_dl_int8_instance.hpp +index 3c4987f..62ee1a4 100644 +--- a/external/composable_kernel/library/src/tensor_operation_instance/gpu/quantization/conv2d_fwd/device_conv2d_dl_int8_instance.hpp ++++ b/external/composable_kernel/library/src/tensor_operation_instance/gpu/quantization/conv2d_fwd/device_conv2d_dl_int8_instance.hpp +@@ -4,7 +4,7 @@ + #pragma once + + #include "conv2d_quantization_common.hpp" +-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" ++#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp" + + namespace ck { + namespace tensor_operation { diff --git a/media-libs/oidn/oidn-2.3.1.ebuild b/media-libs/oidn/oidn-2.3.2.ebuild similarity index 64% rename from media-libs/oidn/oidn-2.3.1.ebuild rename to media-libs/oidn/oidn-2.3.2.ebuild index e461dfc6169f..c8e068d4a987 100644 --- a/media-libs/oidn/oidn-2.3.1.ebuild +++ b/media-libs/oidn/oidn-2.3.2.ebuild @@ -4,16 +4,17 @@ EAPI=8 PYTHON_COMPAT=( python3_{10..13} ) -ROCM_VERSION=5.7 +ROCM_VERSION=6.3 inherit cmake cuda python-any-r1 rocm DESCRIPTION="Intel® Open Image Denoise library" HOMEPAGE="https://www.openimagedenoise.org https://github.com/RenderKit/oidn"; -if [[ ${PV} = *9999 ]]; then +if [[ ${PV} = *9999* ]]; then EGIT_REPO_URI="https://github.com/RenderKit/oidn.git"; EGIT_BRANCH="master" + EGIT_LFS="1" inherit git-r3 else SRC_URI="https://github.com/RenderKit/${PN}/releases/download/v${PV}/${P}.src.tar.gz -> ${P}.tar.gz" @@ -21,18 +22,18 @@ else fi LICENSE="Apache-2.0" -SLOT="0" +SLOT="0/${PV}" +IUSE="apps cuda hip openimageio test" REQUIRED_USE=" test? ( apps ) " -IUSE="apps cuda hip openimageio test" RESTRICT="!test? ( test )" RDEPEND=" dev-cpp/tbb:= dev-lang/ispc - cuda? ( dev-util/nvidia-cuda-toolkit ) - hip? ( dev-util/hip ) + cuda? ( dev-util/nvidia-cuda-toolkit:= ) + hip? ( dev-util/hip:= ) openimageio? ( media-libs/openimageio:= ) " DEPEND="${RDEPEND}" @@ -40,7 +41,6 @@ BDEPEND="${PYTHON_DEPS}" PATCHES=( "${FILESDIR}/${PN}-2.2.2-amdgpu-targets.patch" - "${FILESDIR}/${PN}-2.3.1-hip-clang-19.patch" ) src_prepare() { @@ -49,12 +49,20 @@ src_prepare() { fi if use hip; then - # Fix Clang 19 error - # Bug: https://github.com/RenderKit/oidn/issues/250 - sed -i "s/.template Run(/.template Run<>(/g" \ - external/composable_kernel/include/ck/tensor_operation/gpu/block/blockwise_gemm_wmma.hpp \ - external/composable_kernel/include/ck/tensor_operation/gpu/block/blockwise_gemm_xdlops_skip_b_lds.hpp \ - external/composable_kernel/include/ck/tensor_operation/gpu/block/blockwise_gemm_xdlops.hpp || die + if has_version ">=dev-util/hip-6.2"; then + eapply "${FILESDIR}/${PN}-2.3.1-hip-clang-19.patch" + eapply "${FILESDIR}/${PN}-2.3.1-system-composable-kernel.patch" + eapply "${FILESDIR}/${PN}-2.3.1-composable-kernel-api.patch" + fi + + if has_version "dev-util/hip[llvm_slot_19]"; then + # Fix Clang 19 error + # Bug: https://github.com/RenderKit/oidn/issues/250 + sed -i "s/.template Run(/.template Run<>(/g" \ + external/composable_kernel/include/ck/tensor_operation/gpu/block/blockwise_gemm_wmma.hpp \ + external/composable_kernel/include/ck/tensor_operation/gpu/block/blockwise_gemm_xdlops_skip_b_lds.hpp \ + external/composable_kernel/include/ck/tensor_operation/gpu/block/blockwise_gemm_xdlops.hpp || die + fi # https://bugs.gentoo.org/930391 sed "/-Wno-unused-result/s:): --rocm-path=${EPREFIX}/usr):" \ @@ -65,6 +73,9 @@ src_prepare() { # do not fortify source -- bug 895018 sed -e "s/-D_FORTIFY_SOURCE=2//g" -i {cmake/oidn_platform,external/mkl-dnn/cmake/SDL}.cmake || die + # 950261 + sed -e "/set(CMAKE_CXX_STANDARD /s/^/# /" -i cmake/oidn_platform.cmake || die + cmake_src_prepare } @@ -89,7 +100,7 @@ src_configure() { if use hip; then mycmakeargs+=( -DROCM_PATH="${EPREFIX}/usr" - -DOIDN_DEVICE_HIP_COMPILER="${ESYSROOT}/usr/bin/hipcc" + -DOIDN_DEVICE_HIP_COMPILER="${ESYSROOT}/usr/bin/hipcc" # use HIPHOSTCOMPILER -DAMDGPU_TARGETS="$(get_amdgpu_flags)" ) fi @@ -98,7 +109,7 @@ src_configure() { } src_test() { - "${BUILD_DIR}"/oidnTest || die "There were test faliures!" + "${BUILD_DIR}"/oidnTest || die "There were test failures!" } src_install() {
