Dan Nigrin

Using Max/MSP 4.6 on a MacIntel, I was shocked to find that after 

turning my patch into a standalone application, its size was 30+ MB.

To debug, I made a patch with nothing in it, built it into a 

202 Hack / PC-1600 User / VSTi Host / OMS Convert / Jack OS X

Using Max/MSP 4.6 on a MacIntel, I was shocked to find that after 
turning my patch into a standalone application, its size was 30+ MB.

To debug, I made a patch with nothing in it, built it into a 
standalone, and that is 15.1 MB.

Dan
-- 
Dan Nigrin
Defective Records
202 Hack / PC-1600 User / VSTi Host / OMS Convert / Jack OS X
http://www.defectiverecords.com
http://www.jackosx.com


mega-app-size

Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled' 

software, but a repackaged runtime that loads only your patch? Im 

On Aug 16, 2006, at 4:12 PM, Dan Nigrin wrote:

> Using Max/MSP 4.6 on a MacIntel, I was shocked to find that after 

> turning my patch into a standalone application, its size was 30+ MB.

> To debug, I made a patch with nothing in it, built it into a 

> 202 Hack / PC-1600 User / VSTi Host / OMS Convert / Jack OS X

Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled'  
software, but a repackaged runtime that loads only your patch? Im  
suprised the standalone was 15 megs ;)

> Using Max/MSP 4.6 on a MacIntel, I was shocked to find that after  
> turning my patch into a standalone application, its size was 30+ MB.
>
> To debug, I made a patch with nothing in it, built it into a  
> standalone, and that is 15.1 MB.
>
> Something don't seem right.
>
> Dan
> -- 
> Dan Nigrin
> Defective Records
> 202 Hack / PC-1600 User / VSTi Host / OMS Convert / Jack OS X
> http://www.defectiverecords.com
> http://www.jackosx.com
>

Well, considering that the 4.5.x Runtime was only 2.2 MB, this is a 

real drag if this is the way it's going to have to be... Cycling?

>Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled' 

>software, but a repackaged runtime that loads only your patch? Im 

Well, considering that the 4.5.x Runtime was only 2.2 MB, this is a 
real drag if this is the way it's going to have to be... Cycling?

At 4:30 PM -0400 8/16/06, vade wrote:
>Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled' 
>software, but a repackaged runtime that loads only your patch? Im 
>suprised the standalone was 15 megs ;)

-- 
Dan Nigrin
Defective Records
202 Hack / PC-1600 User / VSTi Host / OMS Convert / Jack OS X
http://www.defectiverecords.com
http://www.jackosx.com


Not knowing very much about how Universal Binary really work, I would 

expect MaxPlay 4.6 to be at least twice as large as Max 4.5 as it needs 

to contain binaries for both PPC and Intel.

> Well, considering that the 4.5.x Runtime was only 2.2 MB, this is a 

> real drag if this is the way it's going to have to be... Cycling?

>> Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled' 

>> software, but a repackaged runtime that loads only your patch? Im 

>> suprised the standalone was 15 megs ;)

Not knowing very much about how Universal Binary really work, I would 
expect MaxPlay 4.6 to be at least twice as large as Max 4.5 as it needs 
to contain binaries for both PPC and Intel.

> Well, considering that the 4.5.x Runtime was only 2.2 MB, this is a 
> real drag if this is the way it's going to have to be... Cycling?
>> Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled' 
>> software, but a repackaged runtime that loads only your patch? Im 
>> suprised the standalone was 15 megs ;)
>


Right Trond, but that's why I was suprised to see the ~5 fold 

At 10:47 PM +0200 8/16/06, Trond Lossius wrote:

>Not knowing very much about how Universal Binary really work, I 

>would expect MaxPlay 4.6 to be at least twice as large as Max 4.5 as 

>it needs to contain binaries for both PPC and Intel.

>>Well, considering that the 4.5.x Runtime was only 2.2 MB, this is a 

>>real drag if this is the way it's going to have to be... Cycling?

>>>Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled' 

>>>software, but a repackaged runtime that loads only your patch? Im 

>>>suprised the standalone was 15 megs ;)

Right Trond, but that's why I was suprised to see the ~5 fold 
increase, not ~2 fold...

At 10:47 PM +0200 8/16/06, Trond Lossius wrote:
>Not knowing very much about how Universal Binary really work, I 
>would expect MaxPlay 4.6 to be at least twice as large as Max 4.5 as 
>it needs to contain binaries for both PPC and Intel.
>
>Trond
>
>>Well, considering that the 4.5.x Runtime was only 2.2 MB, this is a 
>>real drag if this is the way it's going to have to be... Cycling?
>>>Why? The runtime itself for 4.6 is 18 megs. This isnt 'compiled' 
>>>software, but a repackaged runtime that loads only your patch? Im 
>>>suprised the standalone was 15 megs ;)

On Aug 16, 2006, at 1:12 PM, Dan Nigrin wrote:

UB requires much bulkier apps since there is code for both platforms 

in the same executable leading to ~2x the size. If you look inside 

the package you can find out specific file sizes:

Frameworks/MaxAudioAPI.framework: 0.4 MB (can remove if you're not 

Frameworks/MaxJSRef.framework: 0.9 MB (can remove if you're not using 

Frameworks/JitterAPI.framework: 4.1 MB (can remove if you're not 

Resources/MaxMSPCFMSupport.pkg: 0.3MB (can remove if you don't need 

So a minimal UB standalone will be at least 11.5 MB.

> Using Max/MSP 4.6 on a MacIntel, I was shocked to find that after  
> turning my patch into a standalone application, its size was 30+ MB.
>
> To debug, I made a patch with nothing in it, built it into a  
> standalone, and that is 15.1 MB.
>
> Something don't seem right.

UB requires much bulkier apps since there is code for both platforms  
in the same executable leading to ~2x the size.  If you look inside  
the package you can find out specific file sizes:

Frameworks/MaxAPI.framework: 0.7 MB
Frameworks/MaxAudioAPI.framework: 0.4 MB (can remove if you're not  
using MSP)
Frameworks/MaxJSRef.framework: 0.9 MB (can remove if you're not using  
JS)
Frameworks/JitterAPI.framework: 4.1 MB (can remove if you're not  
using Jitter)

Resources/MaxMSPCFMSupport.pkg: 0.3MB (can remove if you don't need  
CFM support)

At 1:58 PM -0700 8/16/06, Joshua Kit Clayton wrote:

>On Aug 16, 2006, at 1:12 PM, Dan Nigrin wrote:

>>Using Max/MSP 4.6 on a MacIntel, I was shocked to find that after 

>>turning my patch into a standalone application, its size was 30+ MB.

>>To debug, I made a patch with nothing in it, built it into a 

>UB requires much bulkier apps since there is code for both platforms 

>in the same executable leading to ~2x the size. If you look inside 

>the package you can find out specific file sizes:

>Frameworks/MaxAudioAPI.framework: 0.4 MB (can remove if you're not using MSP)

>Frameworks/MaxJSRef.framework: 0.9 MB (can remove if you're not using JS)

>Frameworks/JitterAPI.framework: 4.1 MB (can remove if you're not using Jitter)

>Resources/MaxMSPCFMSupport.pkg: 0.3MB (can remove if you don't need 

>So a minimal UB standalone will be at least 11.5 MB.

Thanks Joshua - but I'm still puzzled by the much greater than ~2 

Also, beyond the sizes for the framework and the "repackaged" 

Runtime, my "app" (within the MacOS folder in the package) is now 

10.7 MB, compared to 2 MB when built using Max/MSP 4.5.7 on PPC...

At 1:58 PM -0700 8/16/06, Joshua Kit Clayton wrote:
>On Aug 16, 2006, at 1:12 PM, Dan Nigrin wrote:
>
>>Using Max/MSP 4.6 on a MacIntel, I was shocked to find that after 
>>turning my patch into a standalone application, its size was 30+ MB.
>>
>>To debug, I made a patch with nothing in it, built it into a 
>>standalone, and that is 15.1 MB.
>>
>>Something don't seem right.
>
>UB requires much bulkier apps since there is code for both platforms 
>in the same executable leading to ~2x the size.  If you look inside 
>the package you can find out specific file sizes:
>
>MaxMSPRuntime: 10.8 MB
>
>Frameworks/MaxAPI.framework: 0.7 MB
>Frameworks/MaxAudioAPI.framework: 0.4 MB (can remove if you're not using MSP)
>Frameworks/MaxJSRef.framework: 0.9 MB (can remove if you're not using JS)
>Frameworks/JitterAPI.framework: 4.1 MB (can remove if you're not using Jitter)
>
>Resources/MaxMSPCFMSupport.pkg: 0.3MB (can remove if you don't need 
>CFM support)
>
>So a minimal UB standalone will be at least 11.5 MB.

Thanks Joshua - but I'm still puzzled by the much greater than ~2 
fold increase.

Also, beyond the sizes for the framework and the "repackaged" 
Runtime, my "app" (within the MacOS folder in the package) is now 
10.7 MB, compared to 2 MB when built using Max/MSP 4.5.7 on PPC...

On Aug 16, 2006, at 1:53 PM, Dan Nigrin wrote:

> Right Trond, but that's why I was suprised to see the ~5 fold 

You're neglecting the fact that the Mach-O version uses a different 

compiler and executable format which has the potential to create 

larger executables regardless. This isn't particularly under our 

control or going to change any time soon, so it looks like you have a 

choice between ticking with CFM/PPC, ot upgrading your server bandwidth.

If you recall the quad FAT format of NeXTStep (Motorola, Intel, PA- 

RISC and SPARC), it doesn't seem like binaries are going to get any 

smaller as long as one is compiling machine specific code for 

> Right Trond, but that's why I was suprised to see the ~5 fold  
> increase, not ~2 fold...

You're neglecting the fact that the Mach-O version uses a different  
compiler and executable format which has the potential to create  
larger executables regardless. This isn't particularly under our  
control or going to change any time soon, so it looks like you have a  
choice between ticking with CFM/PPC, ot upgrading your server bandwidth.

If you recall the quad FAT format of NeXTStep (Motorola, Intel, PA- 
RISC and SPARC), it doesn't seem like binaries are going to get any  
smaller as long as one is compiling machine specific code for  
multiple processors.

Maybe you should remove header files from the frameworks copied into 

built application? This is not saving much space, but it's quite useless...

> Frameworks/MaxAudioAPI.framework: 0.4 MB (can remove if you're not using 

> Frameworks/MaxJSRef.framework: 0.9 MB (can remove if you're not using JS)

> Frameworks/JitterAPI.framework: 4.1 MB (can remove if you're not using 

> Resources/MaxMSPCFMSupport.pkg: 0.3MB (can remove if you don't need CFM 

> So a minimal UB standalone will be at least 11.5 MB.

Maybe you should remove header files from the frameworks copied into 
built application? This is not saving much space, but it's quite useless...

> MaxMSPRuntime: 10.8 MB
>
> Frameworks/MaxAPI.framework: 0.7 MB
> Frameworks/MaxAudioAPI.framework: 0.4 MB (can remove if you're not using 
> MSP)
> Frameworks/MaxJSRef.framework: 0.9 MB (can remove if you're not using JS)
> Frameworks/JitterAPI.framework: 4.1 MB (can remove if you're not using 
> Jitter)
>
> Resources/MaxMSPCFMSupport.pkg: 0.3MB (can remove if you don't need CFM 
> support)
>
> So a minimal UB standalone will be at least 11.5 MB.
>
> -Joshua
>
>
>


On Aug 16, 2006, at 2:31 PM, Chris wrote:

> Maybe you should remove header files from the frameworks copied 

> into built application? This is not saving much space, but it's 

These are useful in the case that developers want to use these 

frameworks, so these will in all likelihood stay. Free to delete them 

yourself if you like. Since they are text files they compress quite 

> Maybe you should remove header files from the frameworks copied  
> into built application? This is not saving much space, but it's  
> quite useless...

These are useful in the case that developers want to use these  
frameworks, so these will in all likelihood stay. Free to delete them  
yourself if you like. Since they are text files they compress quite  
nicely for download.

On 16-Aug-2006, at 23:31, Joshua Kit Clayton wrote:

> If you recall the quad FAT format of NeXTStep (Motorola, Intel, PA- 

> RISC and SPARC), it doesn't seem like binaries are going to get any 

> smaller as long as one is compiling machine specific code for 

Plus the last mail I had from ADC indicates that Quad UBs are the 

immediate future for Mac OS: 32 and 64 bit for both Intel and PPC.

I will admit to being a little surprised about seeing 64-bit PPC 

included in the announcement, but it's there in 12pt Lucinda Grande.

I'm looking forward to the first multi-GB Max/MSP standalone. Or 

Peter Castine +--> Litter Power & Litter Bundle for Jitter

Interface Building for |home | chez nous|

Max/MSP Extremely cool |bei uns | i nostri|

On 16-Aug-2006, at 23:31, Joshua Kit Clayton wrote:
> If you recall the quad FAT format of NeXTStep (Motorola, Intel, PA- 
> RISC and SPARC), it doesn't seem like binaries are going to get any  
> smaller as long as one is compiling machine specific code for  
> multiple processors.

Plus the last mail I had from ADC indicates that Quad UBs are the  
immediate future for Mac OS: 32 and 64 bit for both Intel and PPC.

I will admit to being a little surprised about seeing 64-bit PPC  
included in the announcement, but it's there in 12pt Lucinda Grande.

I'm looking forward to the first multi-GB Max/MSP standalone. Or  
maybe not.-)

--------------    http://www.bek.no/~pcastine/Litter/    -------------
Peter Castine             +--> Litter Power & Litter Bundle for Jitter
                                Universal Binaries on the way
iCE:  Sequencing,  Recording &
       Interface  Building  for                   |home    | chez nous|
       Max/MSP   Extremely cool                   |bei uns |  i nostri|
       http://www.dspaudio.com/                   http://www.castine.de

MEGA app size?