30 Jun, 2012, Silenus wrote in the 1st comment:
Votes: 0
I noticed the previous thread made mention of mruby. I was curious about the stability of this and how it performs relative to the main cruby branch (anyone have an idea of what is available in 2.0?).
01 Jul, 2012, Runter wrote in the 2nd comment:
Votes: 0
mruby is very early in development. Still a lot of bugs, but relatively stable. Its main goal is to be competitive with Lua and Javascript implementations for small devices. So the gains you get are in memory profile, compiled binary size, and startup speed. For those reasons it will be good for embedding into other applications. There's projects working on allowing embedding sandboxed mruby interpreters inside of other ruby implementations via gems. It's no surprise that nobody has really been interested in comparing the execution speed of cruby vs other ruby implementations. People will be far more interested in mruby vs lua and javascript. I expect it will be slower than both of these languages for most cases, but still in the ballpark.
01 Jul, 2012, Tyche wrote in the 3rd comment:
Votes: 0
mruby is pretty cool. Not for its speed, but it's "hackability".
Both the yacc code and core support code is about 25% the size of the latest Ruby MRI 1.9.3.

BTW ruby supports goto now!
i = 0
__label__ :loop
puts i
i += 1
__goto__ :loop if i < 10
puts "done"

You have to compile it in though. Do a search for the #define SUPPORT_JOKE and set it to 1
02 Jul, 2012, Silenus wrote in the 4th comment:
Votes: 0
How good is it in terms of stability atm? I have been looking at doing some interpreter hacking.
02 Jul, 2012, Runter wrote in the 5th comment:
Votes: 0
Silenus said:
How good is it in terms of stability atm? I have been looking at doing some interpreter hacking.


I've only played with it so I can't say, but I think it's stable enough that people are already doing interesting things with it. That being said, I wouldn't use it for a for-profit venture until they release more than an "internal team review" version. This is pretty early beta software. How much stability you need is relative. I think they've got any obvious and glaring problems working out simply by using test suites to support the full syntax.
06 Jul, 2012, Silenus wrote in the 6th comment:
Votes: 0
One of the critical aspects of Ruby versus for example Python or C in my opinion- is the lack of good performance (v assembly) comparatively given the level of the language. Obviously devirtualization techniques abound (for example Java- though this language is decidely more static). Closer yet are probably techniques devised for javascript. This is something I am just doing in my spare time and has no particulary monetary value.
10 Jul, 2012, Runter wrote in the 7th comment:
Votes: 0
Silenus said:
One of the critical aspects of Ruby versus for example Python or C in my opinion- is the lack of good performance (v assembly) comparatively given the level of the language. Obviously devirtualization techniques abound (for example Java- though this language is decidely more static). Closer yet are probably techniques devised for javascript. This is something I am just doing in my spare time and has no particulary monetary value.


Well, I think (rightfully) the mruby team have assessed that people pick lua and javascript not because of the syntax or performance (javascript implementations swing wildly!) but because they're the only games in town. A lot of people would love to use something like Ruby in embedded systems, and in Japan mruby is already being embraced by closed shops working on firmware for AV electronics and education centers working in robotics.

So I'm also interested in the benchmarks, but they won't make or break its use. Instead, good enough performance, the full ruby lexicon, and stability will be the determining factors. Compared to a few years ago the budget for such a language is at least 5 times as great, and on some of the sweet spots (like robotics and such) it's factors greater than that. Technology advancing makes performance not the key goal. I think they're looking 5 years down the road and guessing that people will be more interested in object oriented programming in this sector. Well, people already are.
10 Jul, 2012, Tyche wrote in the 8th comment:
Votes: 0
This is an interesting project:
http://code.macournoyer.com/tinyrb/

Interested in benchmarks? It's best to create your own or find code for benchmarks that you think are predictive of your application.
0.0/8