Comments on: 6 Optimization Tips for Ruby MRI

By: Sho Fukamachi Online » Blog Archive » Review: Scaling Ruby by Envycasts/Greg Polack

Thu, 20 Nov 2008 06:27:49 +0000

[...] a couple of minutes using google The “speed tips” at the end are stolen directly from igvita.com, without attribution of course, and the ruby threading graphs look suspiciously similar too. This [...]

By: Compendio de enlaces interesantes #3 | Lobo tuerto

Compendio de enlaces interesantes #3 | Lobo tuerto — Sun, 28 Sep 2008 03:33:48 +0000

[...] sobre el manejo de excepciones en Ruby. Aquí discuten 6 optimizaciones que se pueden aplicar a la hora de hacer programas en [...]

By: Ilya Grigorik

Ilya Grigorik — Mon, 18 Aug 2008 02:25:19 +0000

Rein, I agree, but I also think that context should be taken into account. Readable code is great, but slower code, in a context of extremely performance sensitive application is a poor argument.

By: Rein Henrichs

Rein Henrichs — Thu, 14 Aug 2008 02:19:32 +0000

I think it’s important to keep in mind that optimizing for performanze often pessimizes for the communication value of code. Remember that one of the reasons we use Ruby is for its communicativeness. Performance is merely a factor to be weighed in determining the best way to code something; it is not the only factor or even the most important.

I’d take the inject example out: it’s spurious and an abuse of inject. Also, indirect variable access (creating an accessor for an instance variable, e.g. self.foo) is an OO design decision that has its own merits relative to direct access, even if it is ’slower’. Kent Beck’s Smalltalk Best Practice Patterns has some good things to say on that particular subject.

By: links for 2008-07-17 | Libin Pan

links for 2008-07-17 | Libin Pan — Thu, 17 Jul 2008 06:31:57 +0000

[...] 6 Optimization Tips for Ruby MRI - igvita.com (tags: ruby optimization performance tips programming development rails benchmark) [...]

By: Ilya Grigorik

Ilya Grigorik — Thu, 17 Jul 2008 02:14:15 +0000

Meekish, in a good case, they should be equivalent and cached. In a bad case, instance variables might result in a hash lookup.

James, that’s not skepticism, it’s a legitimate question! For best proof, I would strong recommend and point you to Dr. Stefan Kaes’s ‘Writing Efficient Ruby Code’.

Pistos, great stuff on interpolation!

By: Pistos

Pistos — Wed, 16 Jul 2008 03:26:04 +0000

James H: I’ve benchmarked and published some stats for point 3. I can do the same for point 4 if there is interest.

By: James H

James H — Tue, 15 Jul 2008 15:40:00 +0000

Pardon my skepticism, but do you have any statistics to prove these are more effective?

By: meekish

meekish — Tue, 15 Jul 2008 06:15:17 +0000

Point 2: Isn’t the faster variable an instance variable (not a local variable)?

By: Catholicism Computes » Benchmarking Ruby String interpolation, concatenation and appending

Tue, 15 Jul 2008 04:26:24 +0000

[...] recently came across 6 Optimization Tips for Ruby MRI. One of the tips provided is to use string interpolation instead of concatenation. I set out to [...]