| Message ID | 1551260646-36452-1-git-send-email-Yanqin.Wei@arm.com |
|---|---|
| State | Accepted |
| Commit | d74c920219e3df88a9f535d7e3d2df279463a048 |
| Headers | show |
| Series | [ovs-dev,v1,1/1] hash: Enable hash_bytes128 optimization for aarch64. | expand |
On Wed, Feb 27, 2019 at 05:44:06PM +0800, Yanqin Wei wrote: > "hash_bytes128" has two versions for 64 bits and 32 bits system. This > should be common optimization for their respective platforms. But 64 bits > version was only enabled in x86-64. This patch enable it for aarch64 > platform. > Micro benchmarking test was run in two kinds of arm platform. It was > observed that 50% performance improvement in thunderX2 and 40% improvement > in TaiShan(Cortex-A72). > > Signed-off-by: Yanqin Wei <Yanqin.Wei@arm.com> Thanks, applied to master. (Maybe we should just enable it for all 64-bit systems with something like #if UINTPTR_MAX >= UINT64_MAX.)
diff --git a/lib/hash.c b/lib/hash.c index c64f25e..06f8339 100644 --- a/lib/hash.c +++ b/lib/hash.c @@ -72,7 +72,7 @@ hash_words64__(const uint64_t p[], size_t n_words, uint32_t basis) return hash_words64_inline(p, n_words, basis); } -#if !(defined(__x86_64__)) +#if !(defined(__x86_64__)) && !(defined(__aarch64__)) void hash_bytes128(const void *p_, size_t len, uint32_t basis, ovs_u128 *out) { @@ -233,7 +233,7 @@ hash_bytes128(const void *p_, size_t len, uint32_t basis, ovs_u128 *out) out->u32[3] = h4; } -#else /* __x86_64__ */ +#else /* __x86_64__ or __aarch64__*/ static inline uint64_t hash_rot64(uint64_t x, int8_t r) @@ -361,4 +361,4 @@ hash_bytes128(const void *p_, size_t len, uint32_t basis, ovs_u128 *out) out->u64.lo = h1; out->u64.hi = h2; } -#endif /* __x86_64__ */ +#endif /* __x86_64__ or __aarch64__*/
"hash_bytes128" has two versions for 64 bits and 32 bits system. This should be common optimization for their respective platforms. But 64 bits version was only enabled in x86-64. This patch enable it for aarch64 platform. Micro benchmarking test was run in two kinds of arm platform. It was observed that 50% performance improvement in thunderX2 and 40% improvement in TaiShan(Cortex-A72). Signed-off-by: Yanqin Wei <Yanqin.Wei@arm.com> --- lib/hash.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)