[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PATCH v6 08/10] util/bufferiszero: Simplify test_buffer_is_zero_next_ac
From: |
Richard Henderson |
Subject: |
[PATCH v6 08/10] util/bufferiszero: Simplify test_buffer_is_zero_next_accel |
Date: |
Wed, 24 Apr 2024 15:57:03 -0700 |
Because the three alternatives are monotonic, we don't
need to keep a couple of bitmasks, just identify the
strongest alternative at startup.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
util/bufferiszero.c | 56 ++++++++++++++++++---------------------------
1 file changed, 22 insertions(+), 34 deletions(-)
diff --git a/util/bufferiszero.c b/util/bufferiszero.c
index eb8030a3f0..ff003dc40e 100644
--- a/util/bufferiszero.c
+++ b/util/bufferiszero.c
@@ -179,51 +179,39 @@ buffer_zero_avx2(const void *buf, size_t len)
}
#endif /* CONFIG_AVX2_OPT */
-
-
-static unsigned __attribute__((noinline))
-select_accel_cpuinfo(unsigned info)
-{
- /* Array is sorted in order of algorithm preference. */
- static const struct {
- unsigned bit;
- biz_accel_fn fn;
- } all[] = {
+static biz_accel_fn const accel_table[] = {
+ buffer_is_zero_int_ge256,
+ buffer_zero_sse2,
#ifdef CONFIG_AVX2_OPT
- { CPUINFO_AVX2, buffer_zero_avx2 },
+ buffer_zero_avx2,
#endif
- { CPUINFO_SSE2, buffer_zero_sse2 },
- { CPUINFO_ALWAYS, buffer_is_zero_int_ge256 },
- };
-
- for (unsigned i = 0; i < ARRAY_SIZE(all); ++i) {
- if (info & all[i].bit) {
- buffer_is_zero_accel = all[i].fn;
- return all[i].bit;
- }
- }
- return 0;
-}
-
-static unsigned used_accel;
+};
+static unsigned accel_index;
static void __attribute__((constructor)) init_accel(void)
{
- used_accel = select_accel_cpuinfo(cpuinfo_init());
+ unsigned info = cpuinfo_init();
+ unsigned index = (info & CPUINFO_SSE2 ? 1 : 0);
+
+#ifdef CONFIG_AVX2_OPT
+ if (info & CPUINFO_AVX2) {
+ index = 2;
+ }
+#endif
+
+ accel_index = index;
+ buffer_is_zero_accel = accel_table[index];
}
#define INIT_ACCEL NULL
bool test_buffer_is_zero_next_accel(void)
{
- /*
- * Accumulate the accelerators that we've already tested, and
- * remove them from the set to test this round. We'll get back
- * a zero from select_accel_cpuinfo when there are no more.
- */
- unsigned used = select_accel_cpuinfo(cpuinfo & ~used_accel);
- used_accel |= used;
- return used;
+ if (accel_index != 0) {
+ buffer_is_zero_accel = accel_table[--accel_index];
+ return true;
+ }
+ return false;
}
#else
bool test_buffer_is_zero_next_accel(void)
--
2.34.1
- [PATCH v6 03/10] util/bufferiszero: Reorganize for early test for acceleration, (continued)
- [PATCH v6 03/10] util/bufferiszero: Reorganize for early test for acceleration, Richard Henderson, 2024/04/24
- [PATCH v6 05/10] util/bufferiszero: Optimize SSE2 and AVX2 variants, Richard Henderson, 2024/04/24
- [PATCH v6 06/10] util/bufferiszero: Improve scalar variant, Richard Henderson, 2024/04/24
- [PATCH v6 04/10] util/bufferiszero: Remove useless prefetches, Richard Henderson, 2024/04/24
- [PATCH v6 09/10] util/bufferiszero: Add simd acceleration for aarch64, Richard Henderson, 2024/04/24
- [PATCH v6 10/10] tests/bench: Add bufferiszero-bench, Richard Henderson, 2024/04/24
- [PATCH v6 07/10] util/bufferiszero: Introduce biz_accel_fn typedef, Richard Henderson, 2024/04/24
- [PATCH v6 08/10] util/bufferiszero: Simplify test_buffer_is_zero_next_accel,
Richard Henderson <=