Comments on: Cog http://www.mirandabanda.org/cogblog/2008/06/06/cog/ Speeding Up Croquet and Squeak with a new open-source VM from Teleplace Mon, 05 Sep 2011 20:38:51 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: Alain91 http://www.mirandabanda.org/cogblog/2008/06/06/cog/comment-page-1/#comment-4545 Alain91 Sat, 18 Jun 2011 09:56:30 +0000 http://cogblog.mirandabanda.org/?p=3#comment-4545 Thank you for this kindly answer. Go ahead that's fine. Thank you for this kindly answer. Go ahead that’s fine.

]]> By: Eliot Miranda http://www.mirandabanda.org/cogblog/2008/06/06/cog/comment-page-1/#comment-4544 Eliot Miranda Fri, 17 Jun 2011 22:14:02 +0000 http://cogblog.mirandabanda.org/?p=3#comment-4544 Hi Alain, I think putting effort into a better object representation, that will provide immediate characters, slightly faster message sending and significant;y faster at: and at:put: implementations will go a long way to making the Smalltalk UTF-8 encoding code faster and be more general. So rather than add primitives to the VM for specific VM cases I'd like to put my effort into increasing Smalltalk performance across the board. cheers Eliot Hi Alain,

I think putting effort into a better object representation, that will provide immediate characters, slightly faster message sending and significant;y faster at: and at:put: implementations will go a long way to making the Smalltalk UTF-8 encoding code faster and be more general. So rather than add primitives to the VM for specific VM cases I’d like to put my effort into increasing Smalltalk performance across the board.

cheers
Eliot

]]> By: Alain91 http://www.mirandabanda.org/cogblog/2008/06/06/cog/comment-page-1/#comment-4543 Alain91 Fri, 17 Jun 2011 18:56:33 +0000 http://cogblog.mirandabanda.org/?p=3#comment-4543 I use pharo and seaside and it seems the utf-8 characters are decoded or encoded in smalltalk methods. As utf-8 is now a standard, for performance reason, I think about core primitives in the VM to encode and decode utf-8 character sets. I use pharo and seaside and it seems the utf-8 characters are decoded or encoded in smalltalk methods.
As utf-8 is now a standard, for performance reason, I think about core primitives in the VM to encode and decode utf-8 character sets.

]]>