All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Laight <david.laight.linux@gmail.com>
To: Andrew Morton <akpm@linux-foundation.org>, linux-kernel@vger.kernel.org
Cc: David Laight <david.laight.linux@gmail.com>,
	u.kleine-koenig@baylibre.com, Nicolas Pitre <npitre@baylibre.com>,
	Oleg Nesterov <oleg@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Biju Das <biju.das.jz@bp.renesas.com>
Subject: [PATCH v3 next 09/10] lib: mul_u64_u64_div_u64() Optimise the divide code
Date: Sat, 14 Jun 2025 10:53:45 +0100	[thread overview]
Message-ID: <20250614095346.69130-10-david.laight.linux@gmail.com> (raw)
In-Reply-To: <20250614095346.69130-1-david.laight.linux@gmail.com>

Replace the bit by bit algorithm with one that generates 16 bits
per iteration on 32bit architectures and 32 bits on 64bit ones.

On my zen 5 this reduces the time for the tests (using the generic
code) from ~3350ns to ~1000ns.

Running the 32bit algorithm on 64bit x86 takes ~1500ns.
It'll be slightly slower on a real 32bit system, mostly due
to register pressure.

The savings for 32bit x86 are much higher (tested in userspace).
The worst case (lots of bits in the quotient) drops from ~900 clocks
to ~130 (pretty much independant of the arguments).
Other 32bit architectures may see better savings.

It is possibly to optimise for divisors that span less than
__LONG_WIDTH__/2 bits. However I suspect they don't happen that often
and it doesn't remove any slow cpu divide instructions which dominate
the result.

Signed-off-by: David Laight <david.laight.linux@gmail.com>
---

new patch for v3.

 lib/math/div64.c | 124 +++++++++++++++++++++++++++++++++--------------
 1 file changed, 87 insertions(+), 37 deletions(-)

diff --git a/lib/math/div64.c b/lib/math/div64.c
index fb77fd9d999d..bb318ff2ad87 100644
--- a/lib/math/div64.c
+++ b/lib/math/div64.c
@@ -221,11 +221,37 @@ static inline u64 mul_u64_u64_add_u64(u64 *p_lo, u64 a, u64 b, u64 c)
 
 #endif
 
+#ifndef BITS_PER_ITER
+#define BITS_PER_ITER (__LONG_WIDTH__ >= 64 ? 32 : 16)
+#endif
+
+#if BITS_PER_ITER == 32
+#define mul_u64_long_add_u64(p_lo, a, b, c) mul_u64_u64_add_u64(p_lo, a, b, c)
+#define add_u64_long(a, b) ((a) + (b))
+
+#else
+static inline u32 mul_u64_long_add_u64(u64 *p_lo, u64 a, u32 b, u64 c)
+{
+	u64 n_lo = mul_add(a, b, c);
+	u64 n_med = mul_add(a >> 32, b, c >> 32);
+
+	n_med = add_u64_u32(n_med, n_lo >> 32);
+	*p_lo = n_med << 32 | (u32)n_lo;
+	return n_med >> 32;
+}
+
+#define add_u64_long(a, b) add_u64_u32(a, b)
+
+#endif
+
 u64 mul_u64_add_u64_div_u64(u64 a, u64 b, u64 c, u64 d)
 {
-	u64 n_lo, n_hi;
+	unsigned long d_msig, q_digit;
+	unsigned int reps, d_z_hi;
+	u64 quotient, n_lo, n_hi;
+	u32 overflow;
 
 	if (WARN_ONCE(!d, "%s: division of (%#llx * %#llx + %#llx) by zero, returning 0",
 		      __func__, a, b, c )) {
 		/*
 		 * Return 0 (rather than ~(u64)0) because it is less likely to
@@ -243,46 +269,70 @@ u64 mul_u64_add_u64_div_u64(u64 a, u64 b, u64 c, u64 d)
 		      __func__, a, b, c, n_hi, n_lo, d))
 		return ~(u64)0;
 
-	int shift = __builtin_ctzll(d);
-
-	/* try reducing the fraction in case the dividend becomes <= 64 bits */
-	if ((n_hi >> shift) == 0) {
-		u64 n = shift ? (n_lo >> shift) | (n_hi << (64 - shift)) : n_lo;
-
-		return div64_u64(n, d >> shift);
-		/*
-		 * The remainder value if needed would be:
-		 *   res = div64_u64_rem(n, d >> shift, &rem);
-		 *   rem = (rem << shift) + (n_lo - (n << shift));
-		 */
+	/* Left align the divisor, shifting the dividend to match */
+	d_z_hi = __builtin_clzll(d);
+	if (d_z_hi) {
+		d <<= d_z_hi;
+		n_hi = n_hi << d_z_hi | n_lo >> (64 - d_z_hi);
+		n_lo <<= d_z_hi;
 	}
 
-	/* Do the full 128 by 64 bits division */
-
-	shift = __builtin_clzll(d);
-	d <<= shift;
-
-	int p = 64 + shift;
-	u64 res = 0;
-	bool carry;
+	reps = 64 / BITS_PER_ITER;
+	/* Optimise loop count for small dividends */
+	if (!(u32)(n_hi >> 32)) {
+		reps -= 32 / BITS_PER_ITER;
+		n_hi = n_hi << 32 | n_lo >> 32;
+		n_lo <<= 32;
+	}
+#if BITS_PER_ITER == 16
+	if (!(u32)(n_hi >> 48)) {
+		reps--;
+		n_hi = add_u64_u32(n_hi << 16, n_lo >> 48);
+		n_lo <<= 16;
+	}
+#endif
 
-	do {
-		carry = n_hi >> 63;
-		shift = carry ? 1 : __builtin_clzll(n_hi);
-		if (p < shift)
-			break;
-		p -= shift;
-		n_hi <<= shift;
-		n_hi |= n_lo >> (64 - shift);
-		n_lo <<= shift;
-		if (carry || (n_hi >= d)) {
-			n_hi -= d;
-			res |= 1ULL << p;
+	/* Invert the dividend so we can use add instead of subtract. */
+	n_lo = ~n_lo;
+	n_hi = ~n_hi;
+
+	/*
+	 * Get the most significant BITS_PER_ITER bits of the divisor.
+	 * This is used to get a low 'guestimate' of the quotient digit.
+	 */
+	d_msig = (d >> (64 - BITS_PER_ITER)) + 1;
+
+	/*
+	 * Now do a 'long division' with BITS_PER_ITER bit 'digits'.
+	 * The 'guess' quotient digit can be low and BITS_PER_ITER+1 bits.
+	 * The worst case is dividing ~0 by 0x8000 which requires two subtracts.
+	 */
+	quotient = 0;
+	while (reps--) {
+		q_digit = (unsigned long)(~n_hi >> (64 - 2 * BITS_PER_ITER)) / d_msig;
+		/* Shift 'n' left to align with the product q_digit * d */
+		overflow = n_hi >> (64 - BITS_PER_ITER);
+		n_hi = add_u64_u32(n_hi << BITS_PER_ITER, n_lo >> (64 - BITS_PER_ITER));
+		n_lo <<= BITS_PER_ITER;
+		/* Add product to negated divisor */
+		overflow += mul_u64_long_add_u64(&n_hi, d, q_digit, n_hi);
+		/* Adjust for the q_digit 'guestimate' being low */
+		while (overflow < 0xffffffff >> (32 - BITS_PER_ITER)) {
+			q_digit++;
+			n_hi += d;
+			overflow += n_hi < d;
 		}
-	} while (n_hi);
-	/* The remainder value if needed would be n_hi << p */
+		quotient = add_u64_long(quotient << BITS_PER_ITER, q_digit);
+	}
 
-	return res;
+	/*
+	 * The above only ensures the remainder doesn't overflow,
+	 * it can still be possible to add (aka subtract) another copy
+	 * of the divisor.
+	 */
+	if ((n_hi + d) > n_hi)
+		quotient++;
+	return quotient;
 }
 #if !defined(test_mul_u64_add_u64_div_u64)
 EXPORT_SYMBOL(mul_u64_add_u64_div_u64);
-- 
2.39.5


  parent reply	other threads:[~2025-06-14  9:54 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-14  9:53 [PATCH v3 next 00/10] Implement mul_u64_u64_div_u64_roundup() David Laight
2025-06-14  9:53 ` [PATCH v3 next 01/10] lib: mul_u64_u64_div_u64() rename parameter 'c' to 'd' David Laight
2025-06-14  9:53 ` [PATCH v3 next 02/10] lib: mul_u64_u64_div_u64() Use WARN_ONCE() for divide errors David Laight
2025-06-14 15:17   ` Nicolas Pitre
2025-06-14 21:26     ` David Laight
2025-06-14 22:23       ` Nicolas Pitre
2025-06-14  9:53 ` [PATCH v3 next 03/10] lib: mul_u64_u64_div_u64() simplify check for a 64bit product David Laight
2025-06-14 14:01   ` Nicolas Pitre
2025-06-14  9:53 ` [PATCH v3 next 04/10] lib: Add mul_u64_add_u64_div_u64() and mul_u64_u64_div_u64_roundup() David Laight
2025-06-14 14:06   ` Nicolas Pitre
2025-06-14  9:53 ` [PATCH v3 next 05/10] lib: Add tests for mul_u64_u64_div_u64_roundup() David Laight
2025-06-14 15:19   ` Nicolas Pitre
2025-06-17  4:30   ` Nicolas Pitre
2025-09-18 14:00     ` Uwe Kleine-König
2025-09-18 21:06       ` David Laight
2025-06-14  9:53 ` [PATCH v3 next 06/10] lib: test_mul_u64_u64_div_u64: Test both generic and arch versions David Laight
2025-06-14 15:25   ` Nicolas Pitre
2025-06-18  1:39     ` Nicolas Pitre
2025-06-14  9:53 ` [PATCH v3 next 07/10] lib: mul_u64_u64_div_u64() optimise multiply on 32bit x86 David Laight
2025-06-14 15:31   ` Nicolas Pitre
2025-06-14  9:53 ` [PATCH v3 next 08/10] lib: mul_u64_u64_div_u64() Separate multiply to a helper for clarity David Laight
2025-06-14 15:37   ` Nicolas Pitre
2025-06-14 21:30     ` David Laight
2025-06-14 22:27       ` Nicolas Pitre
2025-06-14  9:53 ` David Laight [this message]
2025-06-17  4:16   ` [PATCH v3 next 09/10] lib: mul_u64_u64_div_u64() Optimise the divide code Nicolas Pitre
2025-06-18  1:33     ` Nicolas Pitre
2025-06-18  9:16       ` David Laight
2025-06-18 15:39         ` Nicolas Pitre
2025-06-18 16:42           ` Nicolas Pitre
2025-06-18 17:54           ` David Laight
2025-06-18 20:12             ` Nicolas Pitre
2025-06-18 22:26               ` David Laight
2025-06-19  2:43                 ` Nicolas Pitre
2025-06-19  8:32                   ` David Laight
2025-06-26 21:46       ` David Laight
2025-06-27  3:48         ` Nicolas Pitre
2025-07-09 14:24   ` David Laight
2025-07-10  9:39     ` 答复: [????] " Li,Rongqing
2025-07-10 10:35       ` David Laight
2025-07-11 21:17     ` David Laight
2025-07-11 21:40       ` Nicolas Pitre
2025-07-14  7:06         ` David Laight
2025-06-14  9:53 ` [PATCH v3 next 10/10] lib: test_mul_u64_u64_div_u64: Test the 32bit code on 64bit David Laight
2025-06-14 10:27 ` [PATCH v3 next 00/10] Implement mul_u64_u64_div_u64_roundup() Peter Zijlstra
2025-06-14 11:59   ` David Laight

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=20250614095346.69130-10-david.laight.linux@gmail.com \
    --to=david.laight.linux@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=biju.das.jz@bp.renesas.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=npitre@baylibre.com \
    --cc=oleg@redhat.com \
    --cc=peterz@infradead.org \
    --cc=u.kleine-koenig@baylibre.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.