[#79440] [Ruby trunk Bug#13188] Reinitialize Ruby VM. — shyouhei@...
Issue #13188 has been updated by Shyouhei Urabe.
6 messages
2017/02/06
[#79441] Re: [Ruby trunk Bug#13188] Reinitialize Ruby VM.
— SASADA Koichi <ko1@...>
2017/02/06
On 2017/02/06 10:10, shyouhei@ruby-lang.org wrote:
[#79532] Immutable Strings vs Symbols — Daniel Ferreira <subtileos@...>
Hi,
15 messages
2017/02/15
[#79541] Re: Immutable Strings vs Symbols
— Rodrigo Rosenfeld Rosas <rr.rosas@...>
2017/02/15
Em 15-02-2017 05:05, Daniel Ferreira escreveu:
[#79543] Re: Immutable Strings vs Symbols
— Daniel Ferreira <subtileos@...>
2017/02/16
Hi Rodrigo,
[#79560] Re: Immutable Strings vs Symbols
— Rodrigo Rosenfeld Rosas <rr.rosas@...>
2017/02/16
Em 15-02-2017 22:39, Daniel Ferreira escreveu:
[ruby-core:79756] [Ruby trunk Feature#13250] Initial estimate for Integer#sqrt should be improved
From:
blogger@...
Date:
2017-02-24 21:36:32 UTC
List:
ruby-core #79756
Issue #13250 has been updated by Nathan Zook. Think about this some more, the fastest solution would be to actually to a N-R round in the floating point unit as well. I will have something in a few days. ---------------------------------------- Feature #13250: Initial estimate for Integer#sqrt should be improved https://bugs.ruby-lang.org/issues/13250#change-63179 * Author: Nathan Zook * Status: Open * Priority: Normal * Assignee: * Target version: ---------------------------------------- r57705, by Nobu, in response to issue #13219, added Integer#sqrt. The initial estimator used is `1 << (b-1)/2 | n >> (b+1)/2`. For this style of estimator, `1 << (b-1)/2 | n >> (b+3)/2` would be best. (Basically, there is a fencepost error in the current commit.) However, much better initial estimates can be made. In particular, using the hardware floating point square root will almost certainly eliminate 3-4 rounds of approximation. Even for smaller numbers, this represents a substantial gain. If for some reason, it is problematic to use fsqrt, a small table lookup can eliminate a couple of rounds. -- https://bugs.ruby-lang.org/ Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe> <http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>