------------------------------------------------------------------- Thu Dec 26 18:08:07 UTC 2024 - scott.bradnick@suse.com - Update to version 0.8.2+git20241219.866cc57: * x86 clmul simde_x_bitreverse_u64: add loongarch implementation (#1249) * x86 avx: reoptimized simde_mm256_addsub_ps/d with lasx * x86 fma: add loongarch lasx optimized implementations * x86 f16c: add loongarch lasx optimized implementations * x86 avx512f: added fmaddsub implementation (#1246) * gh-actions: update CI test for loongarch * x86 sse4.2: add loongarch lsx optimized implementations * x86 sse4.1: add loongarch lsx optimized implementations * x86 ssse3: add loongarch lsx optimized implementations * x86 sse3: add loongarch lsx optimized implementations * x86 sse: add loongarch lsx optimized implementations * x86 avx2: add loongarch lasx optimized implementations (#1241) * x86 avx: add loongarch lasx optimized implementations (#1239) * x86 avx: use INT64_C when the destination is i64 (#1238) * loongarch: add lsx support for sse2.h * x86/avx: Adding overflow test for avx * fix clang type redef error * Whoops, missing comma * gh-actions macos 12: deprecated, going offline on 2024-12-03 * gh-actions msvc-arm64: turn off due to compiler issue * gh-actions macos: xcode 14.3.1 is no longer available, switch to macos-15 to test xcode 16.0 * gh-actions: upgrade Ubuntu Mantic to Ubuntu Noble (24.04) * circleci: switch container for gcc11 i686 -O2 test * gh-actions: pin emsdk to earlier version until https://github.com/llvm/llvm-project/issues/117200 is fixed and released * gh-actions: clean up install list * gh-actions: add loongarch64 clang-18 test * loongarch float16: use a portable version to avoid compilation errors * x86: move definition of 'value' to correct branch in _mm_loadl_epi64 * x86: some better implementations for MSVC and others without SIMDE_STATEMENT_EXPR_ * gh-actions: tests aarch64 without extra features * arm crc32: define SIMDE_ARCH_ARM_CRC32 and consistently use it * _mm256_storeu_pd and _mm256_loadu_pd using 128 bit lanes * gcc-14 -O3 complained about some possible unitialized values * neon/cvz z/Arch: stop using deprecated functions. * gh-actions: upgrade gcc-qemu to gcc-14 * gh-actions: add CI testing for gcc 11 aarch64/arm64 * arm neon: avoid GCC 11 vst1_*_x4 built-in functions * arm neon: fix arm64 gcc11 build excess elements in vector failure * gh-actions: test loongson + lsx with gcc14 from Ubuntu Oracular * x86/sse: Fix type convert error for LSX. * gh-actions: upgrade to Ubuntu 24.04 LTS; upgrade/add GCC 13 / clang 18 * arm wasm: add vst2_u8 translation to Wasm SIMD * arm wasm: add vshll translations to Wasm SIMD * arm wasm: add vst4_u8 translation to Wasm SIMD * appveyor: stop testing with MSVC 2022 until they fix their regressions * math: typo fix, check SIMDE_MATH_NANF instead of the old-style SIMDE_NANF * wasm AltiVec: add u16x8 and u8x16 avgr translations * wasm NEON: add u16x8 and u8x16 avgr translations * arm neon vminnmv_f16: remove duplicate statement (#1208) * x86 wasm: Wasm SIMD version of `_mm_sad_epu8` * msvc: add simde_MemoryBarrier to avoid including * x86/fma: Use 128 bit fnmadd_pd to do 256 bit fnmadd_pd (#1197) * gh-actions: ensure that gcov is present when needed * meson: 0.55.1 is needed for Python 3.12+ * pow: consistently use simde_math_pow * gh-actions: GitHub has retired the macos-11 runners, add some more -13 (x86-64) and -14 (arm64) testing * x86: remove redundant mm_add_pd translation for WASM (#1190) ------------------------------------------------------------------- Mon Jun 24 17:21:42 UTC 2024 - scott.bradnick@suse.com - Update to version 0.8.2+git20240621.c903416: * arm/neon riscv64: additional RVV implementations - part 2. (#1189) * arm/neon riscv64: additional RVV implementations - part1 (#1188) * Use _Float16 in C++ on aarch64 with GCC 13+ * Don't use _Float16 on non-SSE2 x86 * Don't use _Float16 on s390x * x86: Apply half tabular method in _mm_crc32 family * arm: improve performance in vqadd and vmvn in risc-v * neon: avoid warnings when "__ARM_NEON_FP" is not defined. * Add @zengdage as a contributor * Add @howjmay as a contributor * Add @Torinde as a contributor * Add @eric900115 as a contributor * Add @AymenQ as a contributor * add git mail map * start next development cycle: v0.8.3 ------------------------------------------------------------------- Tue Apr 30 16:46:02 UTC 2024 - scott.bradnick@suse.com - Setting "BuildArch: noarch" causes build to fail, commented out - Update to version 0.8.2+git20240430.71fd833: * prepare to release 0.8.0 * arm neon: Fully remove the problematic FCVTZS/FCVTMS/FCVTPS/FCVTNS family intrinsics * arm: improve performance in vabd_xxx for risc-v * arm: improve performance in vhadd_xxx for risc-v * README: two more related projects * arm: Add neon2rvv support in vand series intrinsics * arm: fix some neon2rvv intrinsic function error * arm: enable more intrinsic function for armv7 * wasm x86 impl: some were incorrectly marked SSE instead of SSE2 * arm x86 implementations: allow _m128 access from SSE * tests x86, adjust import order so that MSVC pragma are processed early * svml: SSE is good enough for native m128i and m128d types & functions * sse2 MSVC `_mm_pause` implementaiton for x86 * x86 test msvc: really disable warning 4799,4730 * typo fixes from codespell * README.md - move CLMUL to partial, list more of the CI.yml architectures * Update README.md - link to VPCLMULQDQ; mention MSA (#1157) * Update README.md (#1156) * Packit: Drop i386 (i686) support. (#1155) * gh-actions: build for AMD Buildozer version 2 * x86 xop: fix some native functions * emscripten; use `__builtin_roundeven{f,}` from version 3.1.43 onwards * arm 32 bits: native def fixes; workarounds for gcc * Add gcc arm 32bit armv8-a test in CI * CI: don't run twice on dependabot branches * gh-actions codecov: use token * build(deps): bump codecov/codecov-action from 3 to 4 * update list of fully implemented instruction sets (#1152) * apple clang arm64: ignore SHA2 * semaphore: stop testing on GCC 5 & 6, clang 3.9 & 4 due to forced upgrade to Ubuntu 20.04 * gh-actions macos: report log if there is a configuration failure. * arm platform: cleanup feature detection. * arm neon sm3: check constant range * arm neon: disable some FCVTZS/FCVTMS/FCVTPS/FCVTNS family intrinsics * cirrus: upgrade to clang-17 * gh-actions: test Mac arm64 * arm neon clang: skip vrnd native before clang v18 * clang: detect versions 18 & 19 * build(deps): bump actions/checkout from 3 to 4 (#1149) ------------------------------------------------------------------- Thu Mar 14 14:15:29 UTC 2024 - scott.bradnick@suse.com - Update to version 0.8.0+git20240314.b4e805a: * Initial Support for the RISC-V Vector Extension in ARM NEON (#1130) * start next development cycle: v0.8.1 ------------------------------------------------------------------- Thu Mar 07 16:40:51 UTC 2024 - scott.bradnick@suse.com - Update to version 0.8.0+git20240307.589c7d5: * Add @thomas-schlichter as a contributor * Add @Proudsalsa as a contributor * Add @Coeur as a contributor ------------------------------------------------------------------- Mon Feb 26 17:29:54 UTC 2024 - Jan Engelhardt - Trim unnecessary main package ------------------------------------------------------------------- Mon Feb 26 15:19:39 UTC 2024 - scott.bradnick@suse.com - Update to version 0.8.0-rc1+git20240222.c200c16: * arm/neon abs: negating INT_MIN is undefined behavior * docs: explain how to target a single test * build(deps): bump actions/checkout from 3 to 4 * gh-actions: add new repo for clang-16 * gh-actions emscripten: use older release * docker: ignore common build paths * build(deps): bump github/codeql-action from 2 to 3 (#1138) * build(deps): bump actions/setup-python from 4 to 5 (#1137) * build(deps): bump actions/setup-dotnet from 3 to 4 (#1135) * build(deps): bump ad-m/github-push-action from 0.6.0 to 0.8.0 (#1134) ------------------------------------------------------------------- Mon Feb 26 15:19:21 UTC 2024 - Scott Bradnick - Initial version from upstream.