public inbox for linux-raid@vger.kernel.org
 help / color / mirror / Atom feed
From: "Dirk Müller" <dmueller@suse.de>
To: linux-raid@vger.kernel.org
Cc: "Dirk Müller" <dmueller@suse.de>
Subject: [PATCH] Skip benchmarking of non-best xor_syndrome functions
Date: Wed, 29 Dec 2021 23:34:07 +0100	[thread overview]
Message-ID: <20211229223407.11647-1-dmueller@suse.de> (raw)

In commit fe5cbc6e06c7 ("md/raid6 algorithms: delta syndrome functions")
a xor_syndrome() has been added for optimized syndrome calculation and
being added also to to the raid6_choose_gen() function. However, the
result of the xor_syndrome() benchmarking was intentionally discarded
and did not influence the choice.

We can optimize raid6_choose_gen() without changing its behavior by
only benchmarking the chosen xor_syndrome() variant and printing it
output, rather than benchmarking also the ones that are discarded and
never influence the outcome.

For x86_64, this removes 8 out of 18 benchmark loops which each take
16 jiffies, so up to 160 jiffies saved on module load (640ms on a 250HZ
kernel).

Signed-off-by: Dirk Müller <dmueller@suse.de>
---
 lib/raid6/algos.c | 76 +++++++++++++++++++++++------------------------
 1 file changed, 37 insertions(+), 39 deletions(-)

diff --git a/lib/raid6/algos.c b/lib/raid6/algos.c
index 6d5e5000fdd7..889033b7fc0d 100644
--- a/lib/raid6/algos.c
+++ b/lib/raid6/algos.c
@@ -145,12 +145,12 @@ static inline const struct raid6_recov_calls *raid6_choose_recov(void)
 static inline const struct raid6_calls *raid6_choose_gen(
 	void *(*const dptrs)[RAID6_TEST_DISKS], const int disks)
 {
-	unsigned long perf, bestgenperf, bestxorperf, j0, j1;
+	unsigned long perf, bestgenperf, j0, j1;
 	int start = (disks>>1)-1, stop = disks-3;	/* work on the second half of the disks */
 	const struct raid6_calls *const *algo;
 	const struct raid6_calls *best;
 
-	for (bestgenperf = 0, bestxorperf = 0, best = NULL, algo = raid6_algos; *algo; algo++) {
+	for (bestgenperf = 0, best = NULL, algo = raid6_algos; *algo; algo++) {
 		if (!best || (*algo)->prefer >= best->prefer) {
 			if ((*algo)->valid && !(*algo)->valid())
 				continue;
@@ -180,50 +180,48 @@ static inline const struct raid6_calls *raid6_choose_gen(
 			pr_info("raid6: %-8s gen() %5ld MB/s\n", (*algo)->name,
 				(perf * HZ * (disks-2)) >>
 				(20 - PAGE_SHIFT + RAID6_TIME_JIFFIES_LG2));
+		}
+	}
 
-			if (!(*algo)->xor_syndrome)
-				continue;
+	if (best == NULL) {
+		pr_err("raid6: Yikes! No algorithm found!\n");
+		goto out;
+	}
 
-			perf = 0;
+	raid6_call = *best;
 
-			preempt_disable();
-			j0 = jiffies;
-			while ((j1 = jiffies) == j0)
-				cpu_relax();
-			while (time_before(jiffies,
-					    j1 + (1<<RAID6_TIME_JIFFIES_LG2))) {
-				(*algo)->xor_syndrome(disks, start, stop,
-						      PAGE_SIZE, *dptrs);
-				perf++;
-			}
-			preempt_enable();
-
-			if (best == *algo)
-				bestxorperf = perf;
+	if (!IS_ENABLED(CONFIG_RAID6_PQ_BENCHMARK)) {
+		pr_info("raid6: skipped pq benchmark and selected %s\n",
+			best->name);
+		goto out;
+	}
 
-			pr_info("raid6: %-8s xor() %5ld MB/s\n", (*algo)->name,
-				(perf * HZ * (disks-2)) >>
-				(20 - PAGE_SHIFT + RAID6_TIME_JIFFIES_LG2 + 1));
+	pr_info("raid6: using algorithm %s gen() %ld MB/s\n",
+		best->name,
+		(bestgenperf * HZ * (disks-2)) >>
+		(20 - PAGE_SHIFT+RAID6_TIME_JIFFIES_LG2));
+
+	if (best->xor_syndrome) {
+		perf = 0;
+
+		preempt_disable();
+		j0 = jiffies;
+		while ((j1 = jiffies) == j0)
+			cpu_relax();
+		while (time_before(jiffies,
+				   j1 + (1 << RAID6_TIME_JIFFIES_LG2))) {
+			best->xor_syndrome(disks, start, stop,
+					   PAGE_SIZE, *dptrs);
+			perf++;
 		}
-	}
+		preempt_enable();
 
-	if (best) {
-		if (IS_ENABLED(CONFIG_RAID6_PQ_BENCHMARK)) {
-			pr_info("raid6: using algorithm %s gen() %ld MB/s\n",
-				best->name,
-				(bestgenperf * HZ * (disks-2)) >>
-				(20 - PAGE_SHIFT+RAID6_TIME_JIFFIES_LG2));
-			if (best->xor_syndrome)
-				pr_info("raid6: .... xor() %ld MB/s, rmw enabled\n",
-					(bestxorperf * HZ * (disks-2)) >>
-					(20 - PAGE_SHIFT + RAID6_TIME_JIFFIES_LG2 + 1));
-		} else
-			pr_info("raid6: skip pq benchmark and using algorithm %s\n",
-				best->name);
-		raid6_call = *best;
-	} else
-		pr_err("raid6: Yikes!  No algorithm found!\n");
+		pr_info("raid6: .... xor() %ld MB/s, rmw enabled\n",
+			(perf * HZ * (disks-2)) >>
+			(20 - PAGE_SHIFT + RAID6_TIME_JIFFIES_LG2 + 1));
+	}
 
+out:
 	return best;
 }
 
-- 
2.34.1


             reply	other threads:[~2021-12-29 22:34 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-29 22:34 Dirk Müller [this message]
2021-12-30 11:33 ` [PATCH] Skip benchmarking of non-best xor_syndrome functions Paul Menzel
2021-12-31  8:35   ` Dirk Müller
2021-12-31  8:53     ` Paul Menzel
2022-01-01 23:49 ` Song Liu
2022-01-03 13:49   ` Dirk Müller
2022-01-04 17:40     ` Song Liu

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=20211229223407.11647-1-dmueller@suse.de \
    --to=dmueller@suse.de \
    --cc=linux-raid@vger.kernel.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