From e6366ea640160c6ebb05a7cd6dffd3062ff6bf0f Mon Sep 17 00:00:00 2001 From: nihui Date: Sat, 2 Mar 2024 11:22:46 +0800 Subject: [PATCH] detect arm half (#66) --- .github/workflows/ci.yml | 2 +- README.md | 2 +- main.c | 1 + ruapu.h | 3 +++ 4 files changed, 6 insertions(+), 2 deletions(-) diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 6df7680..2ff89d2 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -182,4 +182,4 @@ jobs: - name: build-test run: | docker run -u root --rm -v $(pwd):/${{ github.workspace }} -w ${{ github.workspace }} --platform linux/${{ matrix.arch }} debian:unstable-slim \ - /bin/bash -c "apt update && apt install -y build-essential && gcc main.c -o ruapu && ./ruapu" + /bin/bash -c "apt update && apt install -y gcc libc6-dev && gcc main.c -o ruapu && ./ruapu" diff --git a/README.md b/README.md index 27607af..72d5c07 100644 --- a/README.md +++ b/README.md @@ -564,7 +564,7 @@ _`fma4` on zen1, ISA in hypervisor, etc._ |CPU|ISA| |:---:|---| |x86|`mmx` `sse` `sse2` `sse3` `ssse3` `sse41` `sse42` `sse4a` `xop` `avx` `f16c` `fma` `fma4` `avx2` `avx512f` `avx512bw` `avx512cd` `avx512dq` `avx512vl` `avx512vnni` `avx512bf16` `avx512ifma` `avx512vbmi` `avx512vbmi2` `avx512fp16` `avx512er` `avx5124fmaps` `avx5124vnniw` `avxvnni` `avxvnniint8` `avxifma`| -|arm|`edsp` `neon` `vfpv4` `idiv`| +|arm|`half` `edsp` `neon` `vfpv4` `idiv`| |aarch64|`neon` `vfpv4` `cpuid` `asimdrdm` `asimdhp` `asimddp` `asimdfhm` `bf16` `i8mm` `mte` `sve` `sve2` `svebf16` `svei8mm` `svef32mm` `pmull` `crc32` `aes` `sha1` `sha2` `sha3` `sha512` `sm3` `sm4` `amx`| |mips|`msa`| |powerpc|`vsx`| diff --git a/main.c b/main.c index 145a5a8..18fa661 100644 --- a/main.c +++ b/main.c @@ -76,6 +76,7 @@ int main() PRINT_ISA_SUPPORT(amx) #elif __arm__ || defined(_M_ARM) + PRINT_ISA_SUPPORT(half) PRINT_ISA_SUPPORT(edsp) PRINT_ISA_SUPPORT(neon) PRINT_ISA_SUPPORT(vfpv4) diff --git a/ruapu.h b/ruapu.h index bcce189..6469ba1 100644 --- a/ruapu.h +++ b/ruapu.h @@ -243,11 +243,13 @@ RUAPU_INSTCODE(amx, 0x00201220) // amx setup #elif __arm__ || defined(_M_ARM) #if __thumb__ +RUAPU_INSTCODE(half, 0xf8bd, 0x0000) // ldrh r0,[sp] RUAPU_INSTCODE(edsp, 0xfb20, 0x0000) // smlad r0,r0,r0,r0 RUAPU_INSTCODE(neon, 0xef00, 0x0d40) // vadd.f32 q0,q0,q0 RUAPU_INSTCODE(vfpv4, 0xeea0, 0x0a00) // vfma.f32 s0,s0,s0 RUAPU_INSTCODE(idiv, 0x2003, 0xfb90, 0xf0f0) // movs r0,#3 + sdiv r0,r0,r0 #else +RUAPU_INSTCODE(half, 0xe1dd00b0) // ldrh r0,[sp] RUAPU_INSTCODE(edsp, 0xe7000010) // smlad r0,r0,r0,r0 RUAPU_INSTCODE(neon, 0xf2000d40) // vadd.f32 q0,q0,q0 RUAPU_INSTCODE(vfpv4, 0xeea00a00) // vfma.f32 s0,s0,s0 @@ -358,6 +360,7 @@ RUAPU_ISAENTRY(sm4) RUAPU_ISAENTRY(amx) #elif __arm__ || defined(_M_ARM) +RUAPU_ISAENTRY(half) RUAPU_ISAENTRY(edsp) RUAPU_ISAENTRY(neon) RUAPU_ISAENTRY(vfpv4)