From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from lists.gentoo.org (pigeon.gentoo.org [208.92.234.80]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by finch.gentoo.org (Postfix) with ESMTPS id 3EEB6158CB7 for ; Tue, 17 Aug 2021 19:05:29 +0000 (UTC) Received: from pigeon.gentoo.org (localhost [127.0.0.1]) by pigeon.gentoo.org (Postfix) with SMTP id 62D98E0871; Tue, 17 Aug 2021 19:05:28 +0000 (UTC) Received: from smtp.gentoo.org (woodpecker.gentoo.org [IPv6:2001:470:ea4a:1:5054:ff:fec7:86e4]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by pigeon.gentoo.org (Postfix) with ESMTPS id 3E6F4E0871 for ; Tue, 17 Aug 2021 19:05:28 +0000 (UTC) Received: from oystercatcher.gentoo.org (oystercatcher.gentoo.org [148.251.78.52]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.gentoo.org (Postfix) with ESMTPS id CA535335D06 for ; Tue, 17 Aug 2021 19:05:26 +0000 (UTC) Received: from localhost.localdomain (localhost [IPv6:::1]) by oystercatcher.gentoo.org (Postfix) with ESMTP id 43F11644 for ; Tue, 17 Aug 2021 19:05:25 +0000 (UTC) From: "Conrad Kostecki" To: gentoo-commits@lists.gentoo.org Content-Transfer-Encoding: 8bit Content-type: text/plain; charset=UTF-8 Reply-To: gentoo-dev@lists.gentoo.org, "Conrad Kostecki" Message-ID: <1629227083.907d9140838786ac15b3509abf409ce2d4d6ba38.conikost@gentoo> Subject: [gentoo-commits] repo/gentoo:master commit in: dev-util/intel-graphics-compiler/ X-VCS-Repository: repo/gentoo X-VCS-Files: dev-util/intel-graphics-compiler/Manifest dev-util/intel-graphics-compiler/intel-graphics-compiler-1.0.8365.ebuild X-VCS-Directories: dev-util/intel-graphics-compiler/ X-VCS-Committer: conikost X-VCS-Committer-Name: Conrad Kostecki X-VCS-Revision: 907d9140838786ac15b3509abf409ce2d4d6ba38 X-VCS-Branch: master Date: Tue, 17 Aug 2021 19:05:25 +0000 (UTC) Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-Id: Gentoo Linux mail X-BeenThere: gentoo-commits@lists.gentoo.org X-Auto-Response-Suppress: DR, RN, NRN, OOF, AutoReply X-Archives-Salt: 3ea8e9ea-2bb2-4bd2-b1dd-540ba562b1a1 X-Archives-Hash: 04600760eeb26da8de5aaea8f355f3ba commit: 907d9140838786ac15b3509abf409ce2d4d6ba38 Author: Conrad Kostecki gentoo org> AuthorDate: Tue Aug 17 18:58:47 2021 +0000 Commit: Conrad Kostecki gentoo org> CommitDate: Tue Aug 17 19:04:43 2021 +0000 URL: https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=907d9140 dev-util/intel-graphics-compiler: bump to version 1.0.8365 Package-Manager: Portage-3.0.20, Repoman-3.0.3 Signed-off-by: Conrad Kostecki gentoo.org> dev-util/intel-graphics-compiler/Manifest | 1 + .../intel-graphics-compiler-1.0.8365.ebuild | 81 ++++++++++++++++++++++ 2 files changed, 82 insertions(+) diff --git a/dev-util/intel-graphics-compiler/Manifest b/dev-util/intel-graphics-compiler/Manifest index 420b17b3c1f..588e100ac8c 100644 --- a/dev-util/intel-graphics-compiler/Manifest +++ b/dev-util/intel-graphics-compiler/Manifest @@ -2,3 +2,4 @@ DIST intel-graphics-compiler-1.0.4111.tar.gz 5781246 BLAKE2B 6fe47422b31b0c73d24 DIST intel-graphics-compiler-1.0.4944.tar.gz 6632351 BLAKE2B a3d72730b8c10fd3810fdb8206d9d92da325a094627b678bb9f8d0941e29c1697994a2d98a1b49780fae62dc421e623dea4c3a7e3061347de5f5b644cfb5e105 SHA512 34fafd5387f3b86bef17511015ff1e8daef75ebdff5d1d14ff2000b58968bb502a9ec1aee988858a2992414f6cdb222e7de19c2d877f7fc8cd84a79417e9a85e DIST intel-graphics-compiler-1.0.5353.tar.gz 6642256 BLAKE2B 34b80b30f33dbd57bcea15852c80cdf8bdbf4630a7904946c7ce84481789e7db42f5216ba14140d445e9b0adbfcfe224468e462ea27768f0e8d31525ce56cde1 SHA512 e9f20e2261c5cb79d365d5d8257e8f4830b2364962c48413cc80154469db34c1c844ae5d576cbce2e3c5e5cbe22c7c3710ebd0a7fe7208e5412e81ce20077c80 DIST intel-graphics-compiler-1.0.8173.tar.gz 7471884 BLAKE2B 2248a8d5f23d4578c5bd04a799af65767916b3705e16a8145149794a2cf16e5a0e16c21669336151a241cc52e7672b05eb892ac4925f2c2d0b22673af4ebf67f SHA512 bebccee0699dea48e98ddcdb320cacacda0cdcd36b68e51d932791756b0a8e2171a78cc1921ca2a092b9fe7dbd7a783d4898ada08c58bf856dd29bfc08334fa0 +DIST intel-graphics-compiler-1.0.8365.tar.gz 7509633 BLAKE2B ffe18d18b409db36bd24215d841daaa36d5055d61b9021e385fe9178351b6a92383839e271a12d1e5017150b9697060fe37f803c5153d2f107813aeb2e7a74cf SHA512 6597afbab2c101d69e6a991b5b065ebbded7ccc6227c0ed82b6b0bd33fec26489c30a0a7a9558617578a767ca6d7382a6ab36144f4c64df742c6bd545ce991e9 diff --git a/dev-util/intel-graphics-compiler/intel-graphics-compiler-1.0.8365.ebuild b/dev-util/intel-graphics-compiler/intel-graphics-compiler-1.0.8365.ebuild new file mode 100644 index 00000000000..b504e2949b3 --- /dev/null +++ b/dev-util/intel-graphics-compiler/intel-graphics-compiler-1.0.8365.ebuild @@ -0,0 +1,81 @@ +# Copyright 1999-2021 Gentoo Authors +# Distributed under the terms of the GNU General Public License v2 + +EAPI=7 + +CMAKE_ECLASS=cmake +PYTHON_COMPAT=( python{3_9,3_10} ) +LLVM_MAX_SLOT=10 + +inherit cmake flag-o-matic llvm python-any-r1 + +DESCRIPTION="LLVM-based OpenCL compiler targetting Intel Gen graphics hardware" +HOMEPAGE="https://github.com/intel/intel-graphics-compiler" +SRC_URI="https://github.com/intel/${PN}/archive/igc-${PV}.tar.gz -> ${P}.tar.gz" +S="${WORKDIR}"/${PN}-igc-${PV} + +LICENSE="MIT" +SLOT="0" +KEYWORDS="~amd64" +IUSE="debug" + +BDEPEND=" + ${PYTHON_DEPS} + >=sys-devel/lld-${LLVM_MAX_SLOT} +" + +# Yes, the last dependency does effectively force the use of llvm-10 +# - there are currently no SLOT=9 ebuilds of opencl-clang with mem2reg +# support. Of course with there being no SLOT=9 ebuilds of vc-intrinsics +# (which we'll need soon as well) at all either we are limited to llvm-10 anyway. +DEPEND=" + <=sys-devel/llvm-${LLVM_MAX_SLOT}.9999:= + <=dev-libs/opencl-clang-${LLVM_MAX_SLOT}.9999:= + >=dev-libs/opencl-clang-12.0.0:= +" + +RDEPEND="${DEPEND}" + +PATCHES=( + "${FILESDIR}/${PN}-1.0.9-no_Werror.patch" + "${FILESDIR}/${PN}-1.0.8173-opencl-clang_version.patch" + "${FILESDIR}/${PN}-1.0.8173-fix-missing-limits.patch" +) + +find_best_llvm_slot() { + local candidate_slot + + # Select the same slot as the best opencl-clang + local ocl_clang_ver="$(best_version -d dev-libs/opencl-clang:=)" + einfo "Selecting ${ocl_clang_ver}" + candidate_slot=$(ver_cut 5 ${ocl_clang_ver}) + + # Sanity check - opencl-clang brings the right LLVM slot as its + # dependency so if this is missing, something is very wrong. + has_version -d sys-devel/llvm:${candidate_slot} || die "LLVM slot matching ${ocl_clang_ver} not found (${candidate_slot})" + + echo ${candidate_slot} +} + +src_configure() { + local llvm_slot=$(find_best_llvm_slot) + einfo "Selecting LLVM slot ${llvm_slot}: $(best_version -d sys-devel/llvm:${llvm_slot})" + local llvm_prefix=$(get_llvm_prefix ${llvm_slot}) + local llvm_ver=$(best_version -d sys-devel/llvm:${llvm_slot}) + llvm_ver=${llvm_ver##*-} + + # Since late March 2020 cmake.eclass does not set -DNDEBUG any more, and the way + # IGC uses this definition causes problems for some users (see Bug #718824). + use debug || append-cppflags -DNDEBUG + + # VectorCompiler needs work, at the moment upstream only supports building vc-intrinsics in place. + local mycmakeargs=( + -DCCLANG_SONAME_VERSION=${llvm_slot} + -DCMAKE_LIBRARY_PATH="${llvm_prefix}"/$(get_libdir) + -DIGC_OPTION__FORCE_SYSTEM_LLVM=ON + -DIGC_PREFERRED_LLVM_VERSION=${llvm_ver} + -DIGC_BUILD__VC_ENABLED=no + ) + + cmake_src_configure +}