[PATCH] tcp_cubic: faster cube root
From: Stephen Hemminger <hidden>
Date: 2007-03-06 22:51:11
Also in:
lkml
The Newton-Raphson method is quadratically convergent so only a small fixed number of steps are necessary. Therefore it is faster to unroll the loop. Since div64_64 is no longer inline it won't cause code explosion. Also fixes a bug that can occur if x^2 was bigger than 32 bits. Signed-off-by: Stephen Hemminger <redacted> --- net/ipv4/tcp_cubic.c | 16 +++++----------- 1 file changed, 5 insertions(+), 11 deletions(-)
--- net-2.6.22.orig/net/ipv4/tcp_cubic.c 2007-03-06 12:24:34.000000000 -0800
+++ net-2.6.22/net/ipv4/tcp_cubic.c 2007-03-06 14:43:37.000000000 -0800@@ -96,23 +96,17 @@ */ static u32 cubic_root(u64 a) { - u32 x, x1; + u64 x; /* Initial estimate is based on: * cbrt(x) = exp(log(x) / 3) */ x = 1u << (fls64(a)/3); - /* - * Iteration based on: - * 2 - * x = ( 2 * x + a / x ) / 3 - * k+1 k k - */ - do { - x1 = x; - x = (2 * x + (uint32_t) div64_64(a, x*x)) / 3; - } while (abs(x1 - x) > 1); + /* converges to 32 bits in 3 iterations */ + x = (2 * x + div64_64(a, x*x)) / 3; + x = (2 * x + div64_64(a, x*x)) / 3; + x = (2 * x + div64_64(a, x*x)) / 3; return x; }