From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: "Philippe Mathieu-Daudé" <philmd@linaro.org>
Subject: [PULL 08/10] util/bufferiszero: Simplify test_buffer_is_zero_next_accel
Date: Fri, 3 May 2024 08:13:12 -0700 [thread overview]
Message-ID: <20240503151314.336357-9-richard.henderson@linaro.org> (raw)
In-Reply-To: <20240503151314.336357-1-richard.henderson@linaro.org>
Because the three alternatives are monotonic, we don't need
to keep a couple of bitmasks, just identify the strongest
alternative at startup.
Generalize test_buffer_is_zero_next_accel and init_accel
by always defining an accel_table array.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
util/bufferiszero.c | 81 ++++++++++++++++++++-------------------------
1 file changed, 35 insertions(+), 46 deletions(-)
diff --git a/util/bufferiszero.c b/util/bufferiszero.c
index f9af7841ba..7218154a13 100644
--- a/util/bufferiszero.c
+++ b/util/bufferiszero.c
@@ -27,7 +27,6 @@
#include "host/cpuinfo.h"
typedef bool (*biz_accel_fn)(const void *, size_t);
-static biz_accel_fn buffer_is_zero_accel;
static bool buffer_is_zero_int_lt256(const void *buf, size_t len)
{
@@ -179,60 +178,35 @@ 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;
- }
+static unsigned best_accel(void)
+{
+ unsigned info = cpuinfo_init();
+
+#ifdef CONFIG_AVX2_OPT
+ if (info & CPUINFO_AVX2) {
+ return 2;
}
- return 0;
+#endif
+ return info & CPUINFO_SSE2 ? 1 : 0;
}
-static unsigned used_accel;
-
-static void __attribute__((constructor)) init_accel(void)
-{
- used_accel = select_accel_cpuinfo(cpuinfo_init());
-}
-
-#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;
-}
#else
-bool test_buffer_is_zero_next_accel(void)
-{
- return false;
-}
-
-#define INIT_ACCEL buffer_is_zero_int_ge256
+#define best_accel() 0
+static biz_accel_fn const accel_table[1] = {
+ buffer_is_zero_int_ge256
+};
#endif
-static biz_accel_fn buffer_is_zero_accel = INIT_ACCEL;
+static biz_accel_fn buffer_is_zero_accel;
+static unsigned accel_index;
bool buffer_is_zero_ool(const void *buf, size_t len)
{
@@ -257,3 +231,18 @@ bool buffer_is_zero_ge256(const void *buf, size_t len)
{
return buffer_is_zero_accel(buf, len);
}
+
+bool test_buffer_is_zero_next_accel(void)
+{
+ if (accel_index != 0) {
+ buffer_is_zero_accel = accel_table[--accel_index];
+ return true;
+ }
+ return false;
+}
+
+static void __attribute__((constructor)) init_accel(void)
+{
+ accel_index = best_accel();
+ buffer_is_zero_accel = accel_table[accel_index];
+}
--
2.34.1
next prev parent reply other threads:[~2024-05-03 15:13 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-03 15:13 [PULL 00/10] bufferiszero improvements Richard Henderson
2024-05-03 15:13 ` [PULL 01/10] util/bufferiszero: Remove SSE4.1 variant Richard Henderson
2024-05-03 15:13 ` [PULL 02/10] util/bufferiszero: Remove AVX512 variant Richard Henderson
2024-05-03 15:13 ` [PULL 03/10] util/bufferiszero: Reorganize for early test for acceleration Richard Henderson
2024-05-03 15:13 ` [PULL 04/10] util/bufferiszero: Remove useless prefetches Richard Henderson
2024-05-03 15:13 ` [PULL 05/10] util/bufferiszero: Optimize SSE2 and AVX2 variants Richard Henderson
2024-05-03 15:13 ` [PULL 06/10] util/bufferiszero: Improve scalar variant Richard Henderson
2024-05-03 15:13 ` [PULL 07/10] util/bufferiszero: Introduce biz_accel_fn typedef Richard Henderson
2024-05-03 15:13 ` Richard Henderson [this message]
2024-05-03 15:13 ` [PULL 09/10] util/bufferiszero: Add simd acceleration for aarch64 Richard Henderson
2024-05-03 15:13 ` [PULL 10/10] tests/bench: Add bufferiszero-bench Richard Henderson
2024-05-03 21:42 ` [PULL 00/10] bufferiszero improvements Richard Henderson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240503151314.336357-9-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=philmd@linaro.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).