diff options
author | Michalis Spyrou <michalis.spyrou@arm.com> | 2021-06-07 14:23:57 +0100 |
---|---|---|
committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-06-23 12:25:50 +0000 |
commit | 20fca524baf99402f742ce38c538f2fd07d5fff9 (patch) | |
tree | b63d98383d1ba22bb3ca59d393e4ab9d47a9c762 /src/core/NEON/kernels/arm_gemm/utils.hpp | |
parent | 1d359279e22874121def2ce4bfdb633d94ea5ade (diff) | |
download | ComputeLibrary-20fca524baf99402f742ce38c538f2fd07d5fff9.tar.gz |
Create core library using high priority operators
A smaller core library is created using a subset of the operators.
Changed the structure of filelist.json in order to include more
information about the kernels and make the selection easier.
Resolves: COMPMID-4514
Change-Id: I079ca7d8e64346174eebdd13b834e1dd4dc36ca2
Signed-off-by: Michalis Spyrou <michalis.spyrou@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5786
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/arm_gemm/utils.hpp')
-rw-r--r-- | src/core/NEON/kernels/arm_gemm/utils.hpp | 50 |
1 files changed, 17 insertions, 33 deletions
diff --git a/src/core/NEON/kernels/arm_gemm/utils.hpp b/src/core/NEON/kernels/arm_gemm/utils.hpp index 1269ef62a6..e648ce2fb5 100644 --- a/src/core/NEON/kernels/arm_gemm/utils.hpp +++ b/src/core/NEON/kernels/arm_gemm/utils.hpp @@ -141,52 +141,36 @@ struct IndirectInputArg { }; namespace utils { -namespace { - -#ifdef __ARM_FEATURE_SVE -template<size_t sz> -inline unsigned long get_vector_length_sz() { - unsigned long v; - - __asm ( - "cntb %0" - : "=r" (v) - ); - - return v / sz; -} - -#define VEC_LEN_SPEC(sz, opcode) template <> inline unsigned long get_vector_length_sz<sz>() { unsigned long v; __asm ( opcode " %0" : "=r" (v)); return v; } - -VEC_LEN_SPEC(8, "cntd") -VEC_LEN_SPEC(4, "cntw") -VEC_LEN_SPEC(2, "cnth") -VEC_LEN_SPEC(1, "cntb") -#endif - -} // anonymous namespace - template <typename T> inline unsigned long get_vector_length() { -#ifdef __ARM_FEATURE_SVE - return get_vector_length_sz<sizeof(T)>(); -#else +#if defined(ARM_COMPUTE_ENABLE_SVE) + uint64_t vl; + + __asm __volatile ( + ".inst 0x0420e3e0\n" // CNTB X0, ALL, MUL #1 + "mov %0, X0\n" + : "=r" (vl) + : + : "x0" + ); + + return vl / sizeof(T); +#else // !defined(ARM_COMPUTE_ENABLE_SVE) return 16 / sizeof(T); -#endif +#endif // defined(ARM_COMPUTE_ENABLE_SVE) } template <typename T> inline unsigned long get_vector_length(VLType vl_type) { switch (vl_type) { -#ifdef __ARM_FEATURE_SVE +#if defined(ARM_COMPUTE_ENABLE_SVE) case VLType::SVE: - return get_vector_length_sz<sizeof(T)>(); -#endif + return get_vector_length<T>(); +#endif // defined(ARM_COMPUTE_ENABLE_SVE) default: return 16 / sizeof(T); } } - } // utils namespace } // arm_gemm namespace |