librelist archives

« back to archive

Shoes performance if not C++

Shoes performance if not C++

From:
Adrian Sampaleanu
Date:
2010-10-16 @ 01:53
Hi,

With the new direction that Shoes is taking, has any thought been given to
the performance of an all Ruby implementation? I know that for non-graphical
apps it wouldn't be an issue, but what about apps trying to do animation, or
other more CPU intensive tasks?

Cheers,
Adrian

Re: [shoes] Shoes performance if not C++

From:
Peter Fitzgibbons
Date:
2010-10-17 @ 14:19
Were not actually using an exclusive Ruby implementation.
A more accurate representation of our refactor is that were using
system libs that have a more generic interface ( and pre-existing Ruby
wrappers). Moving away from Cairo/Pango will help us in the future.

Gtk and Qt are indeed implemented in C/++


Peter Fitzgibbons
Sent from my iPhone

On Oct 15, 2010, at 8:54 PM, Adrian Sampaleanu
<adrian.sampaleanu@gmail.com> wrote:

>
> Hi,
>
> With the new direction that Shoes is taking, has any thought been given 
to the performance of an all Ruby implementation? I know that for 
non-graphical apps it wouldn't be an issue, but what about apps trying to 
do animation, or other more CPU intensive tasks?
>
> Cheers,
> Adrian
>

Re: [shoes] Shoes performance if not C++

From:
Devyn Cairns
Date:
2010-10-17 @ 17:09
Yeah, that's about it.

I really want to know what you guys think about targeting Rubinius.

On Sun, Oct 17, 2010 at 7:19 AM, Peter Fitzgibbons <
peter.fitzgibbons@gmail.com> wrote:

> Were not actually using an exclusive Ruby implementation.
> A more accurate representation of our refactor is that were using
> system libs that have a more generic interface ( and pre-existing Ruby
> wrappers). Moving away from Cairo/Pango will help us in the future.
>
> Gtk and Qt are indeed implemented in C/++
>
>
> Peter Fitzgibbons
> Sent from my iPhone
>
> On Oct 15, 2010, at 8:54 PM, Adrian Sampaleanu
> <adrian.sampaleanu@gmail.com> wrote:
>
> >
> > Hi,
> >
> > With the new direction that Shoes is taking, has any thought been given
> to the performance of an all Ruby implementation? I know that for
> non-graphical apps it wouldn't be an issue, but what about apps trying to do
> animation, or other more CPU intensive tasks?
> >
> > Cheers,
> > Adrian
> >
>



-- 
    ~devyn

Re: [shoes] Shoes performance if not C++

From:
Steve Klabnik
Date:
2010-10-17 @ 17:30
I am all for targeting Rubinius, personally. +1

Re: [shoes] Shoes performance if not C++

From:
Devyn Cairns
Date:
2010-10-17 @ 21:42
Yeah, I think Rubinius + Shoes would be a good match. The only possible
issue is that Rubinius only supports up to 1.8.7's syntax + semantics
(although it does have some of the newer things, like Method#parameters,
although it might have always had that)

On Sun, Oct 17, 2010 at 10:30 AM, Steve Klabnik <steve@steveklabnik.com>wrote:

> I am all for targeting Rubinius, personally. +1
>



-- 
    ~devyn

Re: [shoes] Shoes performance if not C++

From:
Steve Klabnik
Date:
2010-10-17 @ 21:48
Yeah, I don't think it's appropriate right this second. Once Rubinius
is totally up to snuff, I think it's great, but not just yet.

Re: [shoes] Shoes performance if not C++

From:
Devyn Cairns
Date:
2010-10-17 @ 21:50
Okay.

On Sun, Oct 17, 2010 at 2:48 PM, Steve Klabnik <steve@steveklabnik.com>wrote:

> Yeah, I don't think it's appropriate right this second. Once Rubinius
> is totally up to snuff, I think it's great, but not just yet.
>



-- 
    ~devyn

Re: [shoes] Shoes performance if not C++

From:
Rob Britton
Date:
2010-10-18 @ 00:11
Are there versions of Shoes that are only 1.9 compatible?

On Sun, Oct 17, 2010 at 5:50 PM, Devyn Cairns <devyn.cairns@gmail.com> wrote:
> Okay.
> On Sun, Oct 17, 2010 at 2:48 PM, Steve Klabnik <steve@steveklabnik.com>
> wrote:
>>
>> Yeah, I don't think it's appropriate right this second. Once Rubinius
>> is totally up to snuff, I think it's great, but not just yet.
>
>
>
> --
>     ~devyn
>

Re: [shoes] Shoes performance if not C++

From:
Peter Fitzgibbons
Date:
2010-10-18 @ 00:44
Please see The Great Ruby Shoootout
2010<http://programmingzen.com/2010/07/19/the-great-ruby-shootout-july-2010/>

<http://programmingzen.com/2010/07/19/the-great-ruby-shootout-july-2010/>I
think we're on a good path w/ 1.9.2 for the next 6 mo or so

Peter Fitzgibbons
(847) 859-9550
Email: peter.fitzgibbons@gmail.com
IM GTalk: peter.fitzgibbons
IM AOL: peter.fitzgibbons@gmail.com


On Sun, Oct 17, 2010 at 7:11 PM, Rob Britton <rob.s.brit@gmail.com> wrote:

> Are there versions of Shoes that are only 1.9 compatible?
>
> On Sun, Oct 17, 2010 at 5:50 PM, Devyn Cairns <devyn.cairns@gmail.com>
> wrote:
> > Okay.
> > On Sun, Oct 17, 2010 at 2:48 PM, Steve Klabnik <steve@steveklabnik.com>
> > wrote:
> >>
> >> Yeah, I don't think it's appropriate right this second. Once Rubinius
> >> is totally up to snuff, I think it's great, but not just yet.
> >
> >
> >
> > --
> >     ~devyn
> >
>

Re: [shoes] Shoes performance if not C++

From:
Cecil Coupe
Date:
2010-10-18 @ 01:03
The current version of Shoes (policeman) is 1.9.1 only
(1.9.2 with a rakefile patch, but not well tested).

On Sun, 2010-10-17 at 20:11 -0400, Rob Britton wrote:
> Are there versions of Shoes that are only 1.9 compatible?
> 
> On Sun, Oct 17, 2010 at 5:50 PM, Devyn Cairns <devyn.cairns@gmail.com> wrote:
> > Okay.
> > On Sun, Oct 17, 2010 at 2:48 PM, Steve Klabnik <steve@steveklabnik.com>
> > wrote:
> >>
> >> Yeah, I don't think it's appropriate right this second. Once Rubinius
> >> is totally up to snuff, I think it's great, but not just yet.
> >
> >
> >
> > --
> >     ~devyn
> >

Re: [shoes] Shoes performance if not C++

From:
Steve Klabnik
Date:
2010-10-18 @ 01:04
Shoes 3 is ruby 1.9.1, future versions are 1.9.2.
On Oct 17, 2010 8:11 PM, "Rob Britton" <rob.s.brit@gmail.com> wrote:
> Are there versions of Shoes that are only 1.9 compatible?
>
> On Sun, Oct 17, 2010 at 5:50 PM, Devyn Cairns <devyn.cairns@gmail.com>
wrote:
>> Okay.
>> On Sun, Oct 17, 2010 at 2:48 PM, Steve Klabnik <steve@steveklabnik.com>
>> wrote:
>>>
>>> Yeah, I don't think it's appropriate right this second. Once Rubinius
>>> is totally up to snuff, I think it's great, but not just yet.
>>
>>
>>
>> --
>>     ~devyn
>>

Re: [shoes] Shoes performance if not C++

From:
Cecil Coupe
Date:
2010-10-16 @ 02:12
Without a doubt, it's an issue that may be a problem. You have to have
something to measure first, and green/blue/othercolors aren't complete
enough to measure. It might all be for naught.. experiments that fail. I
don't happen to think it's going to be that big an issue, but it's too
soon to say.

Red Shoes has some performance quirks and scalability issues. They might
become worse, or get better. But, with no tests and without a 'new'
Shoes to test against, no one knows.
 


On Fri, 2010-10-15 at 21:53 -0400, Adrian Sampaleanu wrote:
> 
> Hi,
> 
> 
> With the new direction that Shoes is taking, has any thought been
> given to the performance of an all Ruby implementation? I know that
> for non-graphical apps it wouldn't be an issue, but what about apps
> trying to do animation, or other more CPU intensive tasks?
> 
> 
> Cheers,
> Adrian
> 

Re: [shoes] Shoes performance if not C++

From:
Devyn Cairns
Date:
2010-10-16 @ 01:59
You do have a valid point. MRI Ruby is fast enough for animations of medium
complexity on most computers, though. If you need more performance, that's
one of the reasons for Brown Shoes—JRuby is extremely fast.

Hmm, maybe we should also try to target Rubinius? I don't know whether the
GTK+ / Qt bindings work with it, but it could be interesting. Supposedly
it's quite lean and fast with its LLVM compilation and whatnot. Plus, their
philosophy of putting as much as possible in Ruby fits with ours. They would
be the experts to this approach.

On Fri, Oct 15, 2010 at 6:53 PM, Adrian Sampaleanu <
adrian.sampaleanu@gmail.com> wrote:

>
> Hi,
>
> With the new direction that Shoes is taking, has any thought been given to
> the performance of an all Ruby implementation? I know that for non-graphical
> apps it wouldn't be an issue, but what about apps trying to do animation, or
> other more CPU intensive tasks?
>
> Cheers,
> Adrian
>
>


-- 
    ~devyn

Re: [shoes] Shoes performance if not C++

From:
Roger Lovelock
Date:
2010-10-16 @ 02:22
  I don't mean to be offensive - but I see lots of excited discussion 
about the next version of Shoes - when about 1/3 to 1/2 of all potential 
users (as opposed to developers) of Shoes are unable to use the current 
version due to Vista display related issues. Shouldn't that be fixed first?

Roger L

Re: [shoes] Shoes performance if not C++

From:
Steve Klabnik
Date:
2010-10-16 @ 17:35
>  I don't mean to be offensive - but I see lots of excited discussion
> about the next version of Shoes - when about 1/3 to 1/2 of all potential
> users (as opposed to developers) of Shoes are unable to use the current
> version due to Vista display related issues. Shouldn't that be fixed first?

Just to throw my two cents in, I agree that the vista situation is
unfortunate. Not to mention the 10.5 situation. It's just simply that
we don't have anyone that has Vista. I certainly can't afford to buy a
copy of windows, just to run Shoes...

And working on the new versions should actually fix the issue, too.

Re: [shoes] Shoes performance if not C++

From:
Devyn Cairns
Date:
2010-10-16 @ 03:00
Well, Satoshi is the only one working on Windows, and I don't think he has
Vista. It's a problem with Vista itself; Windows 7 doesn't have the issue.

Red Shoes is still technically supported, and I know he'll probably do
whatever he can to help. :D

On Fri, Oct 15, 2010 at 7:22 PM, Roger Lovelock
<rogerlovelock@aanet.com.au>wrote:

>  I don't mean to be offensive - but I see lots of excited discussion
> about the next version of Shoes - when about 1/3 to 1/2 of all potential
> users (as opposed to developers) of Shoes are unable to use the current
> version due to Vista display related issues. Shouldn't that be fixed first?
>
> Roger L
>



-- 
    ~devyn

Re: [shoes] Shoes performance if not C++

From:
Cecil Coupe
Date:
2010-10-16 @ 03:03
Truth is never offensive. 

On Sat, 2010-10-16 at 13:22 +1100, Roger Lovelock wrote:
> I don't mean to be offensive - but I see lots of excited discussion 
> about the next version of Shoes - when about 1/3 to 1/2 of all potential 
> users (as opposed to developers) of Shoes are unable to use the current 
> version due to Vista display related issues. Shouldn't that be fixed first?
> 
> Roger L

Re: [shoes] Shoes performance if not C++

From:
Devyn Cairns
Date:
2010-10-16 @ 03:06
Indeed; responsibility is definitely a priority for us.

On Fri, Oct 15, 2010 at 8:03 PM, Cecil Coupe <ccoupe@cableone.net> wrote:

> Truth is never offensive.
>
> On Sat, 2010-10-16 at 13:22 +1100, Roger Lovelock wrote:
> > I don't mean to be offensive - but I see lots of excited discussion
> > about the next version of Shoes - when about 1/3 to 1/2 of all potential
> > users (as opposed to developers) of Shoes are unable to use the current
> > version due to Vista display related issues. Shouldn't that be fixed
> first?
> >
> > Roger L
>
>
>


-- 
    ~devyn