At the moment, ftload just works with power-of-two tables. Can this  
restriction be suspended? This would make working with ftables more  
flexible.

Saving a ftable of non-power-of-two size with ftsave is no problem.  
Just reading doesn't work.

Thanks very much if someone can do this -

	joachim

------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and 
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today. 
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I'd like to ask this again. Thanks for any reply -

	joachim



Am 21.04.2009 um 20:02 schrieb joachim heintz:

> At the moment, ftload just works with power-of-two tables. Can this
> restriction be suspended? This would make working with ftables more
> flexible.
>
> Saving a ftable of non-power-of-two size with ftsave is no problem.
> Just reading doesn't work.
>
> Thanks very much if someone can do this -
>
> 	joachim
>
> ------------------------------------------------------------------------------
> Stay on top of everything new and different, both inside and
> around Java (TM) technology - register by April 22, and save
> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
> 300 plus technical and hands-on sessions. Register today.
> Use priority code J9JMT32. http://p.sf.net/sfu/p
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I am not sure saving works with ftsave. Anyway I change the code for
ftsave & ftload to allow for NP2 tables. Unfortunately, the k-rate
versions ftload_k etc are not possible to be changed at this point, but
with more investigation it might be OK.

Just watchout that you might have trouble if sizes of allocated tables
change in performance. Use this feature with care.

I'll add this change to CVS now.

Victor
----- Original Message ----- 
From: "joachim heintz" 
To: "Developer discussions" 
Sent: Monday, April 27, 2009 2:56 PM
Subject: Re: [Cs-dev] ftload restriction


> I'd like to ask this again. Thanks for any reply -
>
> joachim
>
>
>
> Am 21.04.2009 um 20:02 schrieb joachim heintz:
>
>> At the moment, ftload just works with power-of-two tables. Can this
>> restriction be suspended? This would make working with ftables more
>> flexible.
>>
>> Saving a ftable of non-power-of-two size with ftsave is no problem.
>> Just reading doesn't work.
>>
>> Thanks very much if someone can do this -
>>
>> joachim
>>
>> ------------------------------------------------------------------------------
>> Stay on top of everything new and different, both inside and
>> around Java (TM) technology - register by April 22, and save
>> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
>> 300 plus technical and hands-on sessions. Register today.
>> Use priority code J9JMT32. http://p.sf.net/sfu/p
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Crystal Reports - New Free Runtime and 30 Day Trial
> Check out the new simplified licensign option that enables unlimited
> royalty-free distribution of the report engine for externally facing
> server and web deployment.
> http://p.sf.net/sfu/businessobjects
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel 


------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Thanks a lot, Victor. The situation I missed the ability of reading  
NP2 tables was at i-pass, so I'm glad having this now. Best -
	joachim



Am 27.04.2009 um 20:54 schrieb victor:

> I am not sure saving works with ftsave. Anyway I change the code for
> ftsave & ftload to allow for NP2 tables. Unfortunately, the k-rate
> versions ftload_k etc are not possible to be changed at this point,  
> but
> with more investigation it might be OK.
>
> Just watchout that you might have trouble if sizes of allocated tables
> change in performance. Use this feature with care.
>
> I'll add this change to CVS now.
>
> Victor
> ----- Original Message -----
> From: "joachim heintz" 
> To: "Developer discussions" 
> Sent: Monday, April 27, 2009 2:56 PM
> Subject: Re: [Cs-dev] ftload restriction
>
>
>> I'd like to ask this again. Thanks for any reply -
>>
>> joachim
>>
>>
>>
>> Am 21.04.2009 um 20:02 schrieb joachim heintz:
>>
>>> At the moment, ftload just works with power-of-two tables. Can this
>>> restriction be suspended? This would make working with ftables more
>>> flexible.
>>>
>>> Saving a ftable of non-power-of-two size with ftsave is no problem.
>>> Just reading doesn't work.
>>>
>>> Thanks very much if someone can do this -
>>>
>>> joachim
>>>
>>> ------------------------------------------------------------------------------
>>> Stay on top of everything new and different, both inside and
>>> around Java (TM) technology - register by April 22, and save
>>> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
>>> 300 plus technical and hands-on sessions. Register today.
>>> Use priority code J9JMT32. http://p.sf.net/sfu/p
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Crystal Reports - New Free Runtime and 30 Day Trial
>> Check out the new simplified licensign option that enables unlimited
>> royalty-free distribution of the report engine for externally  
>> facing
>> server and web deployment.
>> http://p.sf.net/sfu/businessobjects
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Crystal Reports - New Free Runtime and 30 Day Trial
> Check out the new simplified licensign option that enables unlimited
> royalty-free distribution of the report engine for externally  
> facing
> server and web deployment.
> http://p.sf.net/sfu/businessobjects
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>


------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Would it be possible/desirable for us to have means
of querying the float size from some header file? This
would facilitate life for people linking against libcsound64
who are not aware of the need for the token USE_DOUBLE.

Victor 





------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

It would be fine with me to have the SConstruct script generate a "myflt.h"
file or some such file that defines MYFLT.  This header could be included by
sysdep.h and installed with every distribution of Csound.

The only problem I can think off the top of my head would be preventing
Scons from rebuilding everything every time it is run (because the myflt.h
header would change and be included by virtually all targets).  The logic in
Sconstruct could be made "smart" to avoid rewriting the header if it exists
AND the contents are correct for the build options specified.

Anthony

victor wrote on 4/27/09 6:10 PM:

> Would it be possible/desirable for us to have means
> of querying the float size from some header file? This
> would facilitate life for people linking against libcsound64
> who are not aware of the need for the token USE_DOUBLE.


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

> Would it be possible/desirable for us to have means
> of querying the float size from some header file? This
> would facilitate life for people linking against libcsound64
> who are not aware of the need for the token USE_DOUBLE.
>
> Victor
>
>

Easy -- could place in a spare slot in CSOUND structure if that is what
you mean.  And initialised somewhere in the libraried stuff.

==John ff


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I thought of that, but would that not be too late for a program to use? I 
mean,
they would have to check this at compile time I suppose, or is there
a simple run time solution?

I originally thought:

if (csoundGetFloatSize(csound) == 32)
  /* do something */

But that would probably mean having two versions of the same code,
for float and doubles, would it not?


----- Original Message ----- 
From: 
To: "Developer discussions" 
Sent: Tuesday, April 28, 2009 9:18 AM
Subject: Re: [Cs-dev] check for float size


>> Would it be possible/desirable for us to have means
>> of querying the float size from some header file? This
>> would facilitate life for people linking against libcsound64
>> who are not aware of the need for the token USE_DOUBLE.
>>
>> Victor
>>
>>
>
> Easy -- could place in a spare slot in CSOUND structure if that is what
> you mean.  And initialised somewhere in the libraried stuff.
>
> ==John ff
>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel 


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Typically, the client could call this function, set a flag, and then
wherever they communicate with the API, they could convert their data
to the appropriate type and/or type cast it to the appropriate type
before calling a function pointer that was type casted to the
appropriate signature.

It might involve type casting the Csound API function pointers to
named functions with the appropriate signature, e.g.

double **(csoundDoubleGetSpout*)(CSOUND) = (double **(*)(CSOUND *)
dlsym(csoundLibrary, "csoundGetSpout");

double **myDoubleBuffer;
float **myFloatBuffer;
if (csoundIsDouble) {
 myDoubleBuffer = csoundDoubleGetSpout(csound);
} else {
 myFloatBuffer = csoundFloatGetSpout(csound);
}

Hope this helps,
Mike

On 4/28/09, victor  wrote:
> I thought of that, but would that not be too late for a program to use? I
> mean,
> they would have to check this at compile time I suppose, or is there
> a simple run time solution?
>
> I originally thought:
>
> if (csoundGetFloatSize(csound) == 32)
>   /* do something */
>
> But that would probably mean having two versions of the same code,
> for float and doubles, would it not?
>
>
> ----- Original Message -----
> From: 
> To: "Developer discussions" 
> Sent: Tuesday, April 28, 2009 9:18 AM
> Subject: Re: [Cs-dev] check for float size
>
>
>>> Would it be possible/desirable for us to have means
>>> of querying the float size from some header file? This
>>> would facilitate life for people linking against libcsound64
>>> who are not aware of the need for the token USE_DOUBLE.
>>>
>>> Victor
>>>
>>>
>>
>> Easy -- could place in a spare slot in CSOUND structure if that is what
>> you mean.  And initialised somewhere in the libraried stuff.
>>
>> ==John ff
>>
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I was under the impression that you are looking for a compile-time solution
because the doubles and floats libraries usually have different names,
making it unlikely that a program might be linked to either one at run-time.

A run-time solution would probably involve dyld-ing the library after the
program is launched and then using code that could flexibly interface to
either floats or doubles (or two sets of code).  Alternatively, on OS X, you
could weak link to both versions of the library if you keep all of the
interface code in separate files.  One file would include csound.h with
USE_DOUBLE and one without.

Anthony 

victor wrote on 4/28/09 11:21 AM:

> I thought of that, but would that not be too late for a program to use? I
> mean,
> they would have to check this at compile time I suppose, or is there
> a simple run time solution?
> 
> I originally thought:
> 
> if (csoundGetFloatSize(csound) == 32)
> /* do something */
> 
> But that would probably mean having two versions of the same code,
> for float and doubles, would it not?
> 
> 
> ----- Original Message -----
> From: 

>> Easy -- could place in a spare slot in CSOUND structure if that is what
>> you mean.  And initialised somewhere in the libraried stuff.


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

The trouble is twofold: linking to the correctly lib and interpreting
MYFLT correctly. I think a compile-time solution is better since
it does not require special code. Michael's solution would work, but
I think it complicates matters, we should just be able to use MYFLT
directly, with no need for extra conversion or copying.

A recent reason for this is that I am supporting Yann Orlarey's efforts to
create a FAUST output option for Csound and I am trying to find
a neat solution for him to generate code that is ready for 64bit or
32bit csound.

Victor
----- Original Message ----- 
From: "Anthony Kozar" 
To: "New Csound Developer list" 
Sent: Tuesday, April 28, 2009 5:40 PM
Subject: Re: [Cs-dev] check for float size


>I was under the impression that you are looking for a compile-time solution
> because the doubles and floats libraries usually have different names,
> making it unlikely that a program might be linked to either one at 
> run-time.
>
> A run-time solution would probably involve dyld-ing the library after the
> program is launched and then using code that could flexibly interface to
> either floats or doubles (or two sets of code).  Alternatively, on OS X, 
> you
> could weak link to both versions of the library if you keep all of the
> interface code in separate files.  One file would include csound.h with
> USE_DOUBLE and one without.
>
> Anthony
>
> victor wrote on 4/28/09 11:21 AM:
>
>> I thought of that, but would that not be too late for a program to use? I
>> mean,
>> they would have to check this at compile time I suppose, or is there
>> a simple run time solution?
>>
>> I originally thought:
>>
>> if (csoundGetFloatSize(csound) == 32)
>> /* do something */
>>
>> But that would probably mean having two versions of the same code,
>> for float and doubles, would it not?
>>
>>
>> ----- Original Message -----
>> From: 
>
>>> Easy -- could place in a spare slot in CSOUND structure if that is what
>>> you mean.  And initialised somewhere in the libraried stuff.
>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel 


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi Victor,

Just a comment, that's great to hear about support for FAUST output!
If you need any help for testing or otherwise, just let me know!

steven

On Tue, Apr 28, 2009 at 10:41 AM, victor  wrote:
> The trouble is twofold: linking to the correctly lib and interpreting
> MYFLT correctly. I think a compile-time solution is better since
> it does not require special code. Michael's solution would work, but
> I think it complicates matters, we should just be able to use MYFLT
> directly, with no need for extra conversion or copying.
>
> A recent reason for this is that I am supporting Yann Orlarey's efforts to
> create a FAUST output option for Csound and I am trying to find
> a neat solution for him to generate code that is ready for 64bit or
> 32bit csound.
>
> Victor
> ----- Original Message -----
> From: "Anthony Kozar" 
> To: "New Csound Developer list" 
> Sent: Tuesday, April 28, 2009 5:40 PM
> Subject: Re: [Cs-dev] check for float size
>
>
>>I was under the impression that you are looking for a compile-time solution
>> because the doubles and floats libraries usually have different names,
>> making it unlikely that a program might be linked to either one at
>> run-time.
>>
>> A run-time solution would probably involve dyld-ing the library after the
>> program is launched and then using code that could flexibly interface to
>> either floats or doubles (or two sets of code).  Alternatively, on OS X,
>> you
>> could weak link to both versions of the library if you keep all of the
>> interface code in separate files.  One file would include csound.h with
>> USE_DOUBLE and one without.
>>
>> Anthony
>>
>> victor wrote on 4/28/09 11:21 AM:
>>
>>> I thought of that, but would that not be too late for a program to use? I
>>> mean,
>>> they would have to check this at compile time I suppose, or is there
>>> a simple run time solution?
>>>
>>> I originally thought:
>>>
>>> if (csoundGetFloatSize(csound) == 32)
>>> /* do something */
>>>
>>> But that would probably mean having two versions of the same code,
>>> for float and doubles, would it not?
>>>
>>>
>>> ----- Original Message -----
>>> From: 
>>
>>>> Easy -- could place in a spare slot in CSOUND structure if that is what
>>>> you mean.  And initialised somewhere in the libraried stuff.
>>
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

yes, I am very pleased Yann is doing this. I'll keep this list posted.

Victor
----- Original Message ----- 
From: "Steven Yi" 
To: "Developer discussions" 
Sent: Tuesday, April 28, 2009 6:43 PM
Subject: Re: [Cs-dev] check for float size


Hi Victor,

Just a comment, that's great to hear about support for FAUST output!
If you need any help for testing or otherwise, just let me know!

steven

On Tue, Apr 28, 2009 at 10:41 AM, victor  wrote:
> The trouble is twofold: linking to the correctly lib and interpreting
> MYFLT correctly. I think a compile-time solution is better since
> it does not require special code. Michael's solution would work, but
> I think it complicates matters, we should just be able to use MYFLT
> directly, with no need for extra conversion or copying.
>
> A recent reason for this is that I am supporting Yann Orlarey's efforts to
> create a FAUST output option for Csound and I am trying to find
> a neat solution for him to generate code that is ready for 64bit or
> 32bit csound.
>
> Victor
> ----- Original Message -----
> From: "Anthony Kozar" 
> To: "New Csound Developer list" 
> Sent: Tuesday, April 28, 2009 5:40 PM
> Subject: Re: [Cs-dev] check for float size
>
>
>>I was under the impression that you are looking for a compile-time 
>>solution
>> because the doubles and floats libraries usually have different names,
>> making it unlikely that a program might be linked to either one at
>> run-time.
>>
>> A run-time solution would probably involve dyld-ing the library after the
>> program is launched and then using code that could flexibly interface to
>> either floats or doubles (or two sets of code). Alternatively, on OS X,
>> you
>> could weak link to both versions of the library if you keep all of the
>> interface code in separate files. One file would include csound.h with
>> USE_DOUBLE and one without.
>>
>> Anthony
>>
>> victor wrote on 4/28/09 11:21 AM:
>>
>>> I thought of that, but would that not be too late for a program to use? 
>>> I
>>> mean,
>>> they would have to check this at compile time I suppose, or is there
>>> a simple run time solution?
>>>
>>> I originally thought:
>>>
>>> if (csoundGetFloatSize(csound) == 32)
>>> /* do something */
>>>
>>> But that would probably mean having two versions of the same code,
>>> for float and doubles, would it not?
>>>
>>>
>>> ----- Original Message -----
>>> From: 
>>
>>>> Easy -- could place in a spare slot in CSOUND structure if that is what
>>>> you mean. And initialised somewhere in the libraried stuff.
>>
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations
Conference from O'Reilly Media. Velocity features a full day of
expert-led, hands-on workshops and two days of sessions from industry
leaders in dedicated Performance & Operations tracks. Use code vel09scf
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-devel 


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I think this is great, I'm looking forward to trying it. I think it
will help us design great Csound instruments; you can essentially
design your instrument as an opcode and pass all the instr pfields to
the opcode.

Regards,
Mike

On 4/28/09, victor  wrote:
> yes, I am very pleased Yann is doing this. I'll keep this list posted.
>
> Victor
> ----- Original Message -----
> From: "Steven Yi" 
> To: "Developer discussions" 
> Sent: Tuesday, April 28, 2009 6:43 PM
> Subject: Re: [Cs-dev] check for float size
>
>
> Hi Victor,
>
> Just a comment, that's great to hear about support for FAUST output!
> If you need any help for testing or otherwise, just let me know!
>
> steven
>
> On Tue, Apr 28, 2009 at 10:41 AM, victor  wrote:
>> The trouble is twofold: linking to the correctly lib and interpreting
>> MYFLT correctly. I think a compile-time solution is better since
>> it does not require special code. Michael's solution would work, but
>> I think it complicates matters, we should just be able to use MYFLT
>> directly, with no need for extra conversion or copying.
>>
>> A recent reason for this is that I am supporting Yann Orlarey's efforts to
>> create a FAUST output option for Csound and I am trying to find
>> a neat solution for him to generate code that is ready for 64bit or
>> 32bit csound.
>>
>> Victor
>> ----- Original Message -----
>> From: "Anthony Kozar" 
>> To: "New Csound Developer list" 
>> Sent: Tuesday, April 28, 2009 5:40 PM
>> Subject: Re: [Cs-dev] check for float size
>>
>>
>>>I was under the impression that you are looking for a compile-time
>>>solution
>>> because the doubles and floats libraries usually have different names,
>>> making it unlikely that a program might be linked to either one at
>>> run-time.
>>>
>>> A run-time solution would probably involve dyld-ing the library after the
>>> program is launched and then using code that could flexibly interface to
>>> either floats or doubles (or two sets of code). Alternatively, on OS X,
>>> you
>>> could weak link to both versions of the library if you keep all of the
>>> interface code in separate files. One file would include csound.h with
>>> USE_DOUBLE and one without.
>>>
>>> Anthony
>>>
>>> victor wrote on 4/28/09 11:21 AM:
>>>
>>>> I thought of that, but would that not be too late for a program to use?
>>>> I
>>>> mean,
>>>> they would have to check this at compile time I suppose, or is there
>>>> a simple run time solution?
>>>>
>>>> I originally thought:
>>>>
>>>> if (csoundGetFloatSize(csound) == 32)
>>>> /* do something */
>>>>
>>>> But that would probably mean having two versions of the same code,
>>>> for float and doubles, would it not?
>>>>
>>>>
>>>> ----- Original Message -----
>>>> From: 
>>>
>>>>> Easy -- could place in a spare slot in CSOUND structure if that is what
>>>>> you mean. And initialised somewhere in the libraried stuff.
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Register Now & Save for Velocity, the Web Performance & Operations
>>> Conference from O'Reilly Media. Velocity features a full day of
>>> expert-led, hands-on workshops and two days of sessions from industry
>>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

My previous suggestion should help eliminate the need to define USE_DOUBLE
in client code.  However, it still requires that the client build script
know where the Csound headers are and what the library is called for the
version that it wants to use.

The build script for a client could compile a simple program to determine
the MYFLT size by using the existing API function csoundGetSizeOfMYFLT() but
compiling that program might have the same issues as compiling the client.

Problems like this are usually solved with other libraries by installing a
program-config script along with the library that helps you find out the
correct compiler and linker options to use.  So, csound-config might output
something like this on OS X for floats

"-I/Library/Frameworks/CsoundLib.framework/Versions/5.2/Headers -framework
CsoundLib"

and this for doubles

"-DUSE_DOUBLE 
-I/Library/Frameworks/CsoundLib64.framework/Versions/5.2/Headers -framework
CsoundLib64"

I'm not sure how we want to proceed though, especially since we intend for
both the floats and doubles versions to be able to coexist on the same
machine.  In the case of FAUST, I suppose using floats or doubles could be a
config option for the build script of the FAUST output ?

Anthony

victor wrote on 4/28/09 1:41 PM:

> The trouble is twofold: linking to the correctly lib and interpreting
> MYFLT correctly. I think a compile-time solution is better since
> it does not require special code. Michael's solution would work, but
> I think it complicates matters, we should just be able to use MYFLT
> directly, with no need for extra conversion or copying.
> 
> A recent reason for this is that I am supporting Yann Orlarey's efforts to
> create a FAUST output option for Csound and I am trying to find
> a neat solution for him to generate code that is ready for 64bit or
> 32bit csound.


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I beg your indulgence as I repeat my call to omit the float version of
csound, and to produce only the double version.

Alternatively, to omit the double version, and produce only the float version.

The criteria would be, of course, to measure how fast each build runs,
and do double-blind listening and mathematical comparison tests to see
how good each build sounds.

The current situation is a big waste of time unless, and only unless,
one of the builds is hopeless at efficiency, and the other is hopeless
at precision.

Regards,
Mike

On 4/28/09, Anthony Kozar  wrote:
> My previous suggestion should help eliminate the need to define USE_DOUBLE
> in client code.  However, it still requires that the client build script
> know where the Csound headers are and what the library is called for the
> version that it wants to use.
>
> The build script for a client could compile a simple program to determine
> the MYFLT size by using the existing API function csoundGetSizeOfMYFLT() but
> compiling that program might have the same issues as compiling the client.
>
> Problems like this are usually solved with other libraries by installing a
> program-config script along with the library that helps you find out the
> correct compiler and linker options to use.  So, csound-config might output
> something like this on OS X for floats
>
> "-I/Library/Frameworks/CsoundLib.framework/Versions/5.2/Headers -framework
> CsoundLib"
>
> and this for doubles
>
> "-DUSE_DOUBLE
> -I/Library/Frameworks/CsoundLib64.framework/Versions/5.2/Headers -framework
> CsoundLib64"
>
> I'm not sure how we want to proceed though, especially since we intend for
> both the floats and doubles versions to be able to coexist on the same
> machine.  In the case of FAUST, I suppose using floats or doubles could be a
> config option for the build script of the FAUST output ?
>
> Anthony
>
> victor wrote on 4/28/09 1:41 PM:
>
>> The trouble is twofold: linking to the correctly lib and interpreting
>> MYFLT correctly. I think a compile-time solution is better since
>> it does not require special code. Michael's solution would work, but
>> I think it complicates matters, we should just be able to use MYFLT
>> directly, with no need for extra conversion or copying.
>>
>> A recent reason for this is that I am supporting Yann Orlarey's efforts to
>> create a FAUST output option for Csound and I am trying to find
>> a neat solution for him to generate code that is ready for 64bit or
>> 32bit csound.
>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Well, I'm all for double version.  However, we have backwards
compatibility to deal with, as some projects will only render with
float version if their analysis files are float.

On Tue, Apr 28, 2009 at 1:02 PM, Michael Gogins
 wrote:
> I beg your indulgence as I repeat my call to omit the float version of
> csound, and to produce only the double version.
>
> Alternatively, to omit the double version, and produce only the float version.
>
> The criteria would be, of course, to measure how fast each build runs,
> and do double-blind listening and mathematical comparison tests to see
> how good each build sounds.
>
> The current situation is a big waste of time unless, and only unless,
> one of the builds is hopeless at efficiency, and the other is hopeless
> at precision.
>
> Regards,
> Mike
>
> On 4/28/09, Anthony Kozar  wrote:
>> My previous suggestion should help eliminate the need to define USE_DOUBLE
>> in client code.  However, it still requires that the client build script
>> know where the Csound headers are and what the library is called for the
>> version that it wants to use.
>>
>> The build script for a client could compile a simple program to determine
>> the MYFLT size by using the existing API function csoundGetSizeOfMYFLT() but
>> compiling that program might have the same issues as compiling the client.
>>
>> Problems like this are usually solved with other libraries by installing a
>> program-config script along with the library that helps you find out the
>> correct compiler and linker options to use.  So, csound-config might output
>> something like this on OS X for floats
>>
>> "-I/Library/Frameworks/CsoundLib.framework/Versions/5.2/Headers -framework
>> CsoundLib"
>>
>> and this for doubles
>>
>> "-DUSE_DOUBLE
>> -I/Library/Frameworks/CsoundLib64.framework/Versions/5.2/Headers -framework
>> CsoundLib64"
>>
>> I'm not sure how we want to proceed though, especially since we intend for
>> both the floats and doubles versions to be able to coexist on the same
>> machine.  In the case of FAUST, I suppose using floats or doubles could be a
>> config option for the build script of the FAUST output ?
>>
>> Anthony
>>
>> victor wrote on 4/28/09 1:41 PM:
>>
>>> The trouble is twofold: linking to the correctly lib and interpreting
>>> MYFLT correctly. I think a compile-time solution is better since
>>> it does not require special code. Michael's solution would work, but
>>> I think it complicates matters, we should just be able to use MYFLT
>>> directly, with no need for extra conversion or copying.
>>>
>>> A recent reason for this is that I am supporting Yann Orlarey's efforts to
>>> create a FAUST output option for Csound and I am trying to find
>>> a neat solution for him to generate code that is ready for 64bit or
>>> 32bit csound.
>>
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
>
> --
> Michael Gogins
> Irreducible Productions
> http://www.michael-gogins.com
> Michael dot Gogins at gmail dot com
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I think that could be dealt with using either conversion utilities, or
some extra coding in the existing utilities.

Regards,
Mike

On 4/28/09, Steven Yi  wrote:
> Well, I'm all for double version.  However, we have backwards
> compatibility to deal with, as some projects will only render with
> float version if their analysis files are float.
>
> On Tue, Apr 28, 2009 at 1:02 PM, Michael Gogins
>  wrote:
>> I beg your indulgence as I repeat my call to omit the float version of
>> csound, and to produce only the double version.
>>
>> Alternatively, to omit the double version, and produce only the float
>> version.
>>
>> The criteria would be, of course, to measure how fast each build runs,
>> and do double-blind listening and mathematical comparison tests to see
>> how good each build sounds.
>>
>> The current situation is a big waste of time unless, and only unless,
>> one of the builds is hopeless at efficiency, and the other is hopeless
>> at precision.
>>
>> Regards,
>> Mike
>>
>> On 4/28/09, Anthony Kozar  wrote:
>>> My previous suggestion should help eliminate the need to define
>>> USE_DOUBLE
>>> in client code.  However, it still requires that the client build script
>>> know where the Csound headers are and what the library is called for the
>>> version that it wants to use.
>>>
>>> The build script for a client could compile a simple program to determine
>>> the MYFLT size by using the existing API function csoundGetSizeOfMYFLT()
>>> but
>>> compiling that program might have the same issues as compiling the
>>> client.
>>>
>>> Problems like this are usually solved with other libraries by installing
>>> a
>>> program-config script along with the library that helps you find out the
>>> correct compiler and linker options to use.  So, csound-config might
>>> output
>>> something like this on OS X for floats
>>>
>>> "-I/Library/Frameworks/CsoundLib.framework/Versions/5.2/Headers
>>> -framework
>>> CsoundLib"
>>>
>>> and this for doubles
>>>
>>> "-DUSE_DOUBLE
>>> -I/Library/Frameworks/CsoundLib64.framework/Versions/5.2/Headers
>>> -framework
>>> CsoundLib64"
>>>
>>> I'm not sure how we want to proceed though, especially since we intend
>>> for
>>> both the floats and doubles versions to be able to coexist on the same
>>> machine.  In the case of FAUST, I suppose using floats or doubles could
>>> be a
>>> config option for the build script of the FAUST output ?
>>>
>>> Anthony
>>>
>>> victor wrote on 4/28/09 1:41 PM:
>>>
>>>> The trouble is twofold: linking to the correctly lib and interpreting
>>>> MYFLT correctly. I think a compile-time solution is better since
>>>> it does not require special code. Michael's solution would work, but
>>>> I think it complicates matters, we should just be able to use MYFLT
>>>> directly, with no need for extra conversion or copying.
>>>>
>>>> A recent reason for this is that I am supporting Yann Orlarey's efforts
>>>> to
>>>> create a FAUST output option for Csound and I am trying to find
>>>> a neat solution for him to generate code that is ready for 64bit or
>>>> 32bit csound.
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Register Now & Save for Velocity, the Web Performance & Operations
>>> Conference from O'Reilly Media. Velocity features a full day of
>>> expert-led, hands-on workshops and two days of sessions from industry
>>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>>
>> --
>> Michael Gogins
>> Irreducible Productions
>> http://www.michael-gogins.com
>> Michael dot Gogins at gmail dot com
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Thank you. On Linux, then, the decision would seem not to require any brains.

I will perform similar tests myself on both Linux and Windows, when I
get the time, and obviously this kind of thing does not actually take
a lot of time to do.

Regards,
Mike

On 4/28/09, Felipe Sateler  wrote:
> El 29/04/09 06:02 Michael Gogins escribió:
>> I beg your indulgence as I repeat my call to omit the float version of
>> csound, and to produce only the double version.
>>
>> Alternatively, to omit the double version, and produce only the float
>> version.
>>
>> The criteria would be, of course, to measure how fast each build runs,
>> and do double-blind listening and mathematical comparison tests to see
>> how good each build sounds.
>>
>> The current situation is a big waste of time unless, and only unless,
>> one of the builds is hopeless at efficiency, and the other is hopeless
>> at precision.
>>
>
> In 64bit linux, doubles seem to be even faster.
>
> http://www.nabble.com/Re:-Re:-Re:-64-bit---double-question-p17329282.html
>
> Saludos,
> Felipe Sateler
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

The floats build will still be of use for older machines. Will the
floats option be completely removed, or the idea is to distribute only
the doubles version, but have the floats version as an alternative
compile option?

Cheers,
Andrés

On Wed, Apr 29, 2009 at 7:19 AM, Michael Gogins
 wrote:
> Thank you. On Linux, then, the decision would seem not to require any brains.
>
> I will perform similar tests myself on both Linux and Windows, when I
> get the time, and obviously this kind of thing does not actually take
> a lot of time to do.
>
> Regards,
> Mike
>
> On 4/28/09, Felipe Sateler  wrote:
>> El 29/04/09 06:02 Michael Gogins escribió:
>>> I beg your indulgence as I repeat my call to omit the float version of
>>> csound, and to produce only the double version.
>>>
>>> Alternatively, to omit the double version, and produce only the float
>>> version.
>>>
>>> The criteria would be, of course, to measure how fast each build runs,
>>> and do double-blind listening and mathematical comparison tests to see
>>> how good each build sounds.
>>>
>>> The current situation is a big waste of time unless, and only unless,
>>> one of the builds is hopeless at efficiency, and the other is hopeless
>>> at precision.
>>>
>>
>> In 64bit linux, doubles seem to be even faster.
>>
>> http://www.nabble.com/Re:-Re:-Re:-64-bit---double-question-p17329282.html
>>
>> Saludos,
>> Felipe Sateler
>>
>
>
> --
> Michael Gogins
> Irreducible Productions
> http://www.michael-gogins.com
> Michael dot Gogins at gmail dot com
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>



-- 


Andrés

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I propose to deprecate the float version, to distribute only the
double version, and to keep the float version around in case anyone
really needs it.

My results are that the float version is somewhat faster on Windows, I
hear that the double version is quite a bit faster on Linux, and I
would like to know what the story is on the Mac.

I do not wish to do anything hasty or that would alienate users, but I
think the current situation is not exactly optimal.

Regards,
Mike

On 4/29/09, Andres Cabrera  wrote:
> The floats build will still be of use for older machines. Will the
> floats option be completely removed, or the idea is to distribute only
> the doubles version, but have the floats version as an alternative
> compile option?
>
> Cheers,
> Andrés
>
> On Wed, Apr 29, 2009 at 7:19 AM, Michael Gogins
>  wrote:
>> Thank you. On Linux, then, the decision would seem not to require any
>> brains.
>>
>> I will perform similar tests myself on both Linux and Windows, when I
>> get the time, and obviously this kind of thing does not actually take
>> a lot of time to do.
>>
>> Regards,
>> Mike
>>
>> On 4/28/09, Felipe Sateler  wrote:
>>> El 29/04/09 06:02 Michael Gogins escribió:
>>>> I beg your indulgence as I repeat my call to omit the float version of
>>>> csound, and to produce only the double version.
>>>>
>>>> Alternatively, to omit the double version, and produce only the float
>>>> version.
>>>>
>>>> The criteria would be, of course, to measure how fast each build runs,
>>>> and do double-blind listening and mathematical comparison tests to see
>>>> how good each build sounds.
>>>>
>>>> The current situation is a big waste of time unless, and only unless,
>>>> one of the builds is hopeless at efficiency, and the other is hopeless
>>>> at precision.
>>>>
>>>
>>> In 64bit linux, doubles seem to be even faster.
>>>
>>> http://www.nabble.com/Re:-Re:-Re:-64-bit---double-question-p17329282.html
>>>
>>> Saludos,
>>> Felipe Sateler
>>>
>>
>>
>> --
>> Michael Gogins
>> Irreducible Productions
>> http://www.michael-gogins.com
>> Michael dot Gogins at gmail dot com
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
>
>
> --
>
>
> Andrés
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I would prefer if this is done by stages. First move USE_DOUBLE to 
USE_FLOAT, so
that MYFLT is by default double in the next release, but still distribute 
both, encouraging
Csound API users to link to 64bit version by default, as well as users .

This can also allow us to assess any outstanding issues that might arise 
with say the
Python module on OSX etc.. I would be a bit worried to drop and move to 
64bit-only
as my experience is that the float version has been probably the most used 
and
debugged (in terms of package release etc), esp. in platforms such as OSX. 
On
fedora, for instance official packages only support floats AFAIK. I am not 
keen
on having a release out there that is problematic and broken in many places.

The transition would allow us to make sure most things are correct. But I 
agree we should
move to 64bit and start the move now. This would be a good thing.

Victor


----- Original Message ----- 
From: "Michael Gogins" 
To: "Developer discussions" 
Sent: Wednesday, April 29, 2009 3:40 PM
Subject: Re: [Cs-dev] check for float size


I propose to deprecate the float version, to distribute only the
double version, and to keep the float version around in case anyone
really needs it.

My results are that the float version is somewhat faster on Windows, I
hear that the double version is quite a bit faster on Linux, and I
would like to know what the story is on the Mac.

I do not wish to do anything hasty or that would alienate users, but I
think the current situation is not exactly optimal.

Regards,
Mike

On 4/29/09, Andres Cabrera  wrote:
> The floats build will still be of use for older machines. Will the
> floats option be completely removed, or the idea is to distribute only
> the doubles version, but have the floats version as an alternative
> compile option?
>
> Cheers,
> Andrés
>
> On Wed, Apr 29, 2009 at 7:19 AM, Michael Gogins
>  wrote:
>> Thank you. On Linux, then, the decision would seem not to require any
>> brains.
>>
>> I will perform similar tests myself on both Linux and Windows, when I
>> get the time, and obviously this kind of thing does not actually take
>> a lot of time to do.
>>
>> Regards,
>> Mike
>>
>> On 4/28/09, Felipe Sateler  wrote:
>>> El 29/04/09 06:02 Michael Gogins escribió:
>>>> I beg your indulgence as I repeat my call to omit the float version of
>>>> csound, and to produce only the double version.
>>>>
>>>> Alternatively, to omit the double version, and produce only the float
>>>> version.
>>>>
>>>> The criteria would be, of course, to measure how fast each build runs,
>>>> and do double-blind listening and mathematical comparison tests to see
>>>> how good each build sounds.
>>>>
>>>> The current situation is a big waste of time unless, and only unless,
>>>> one of the builds is hopeless at efficiency, and the other is hopeless
>>>> at precision.
>>>>
>>>
>>> In 64bit linux, doubles seem to be even faster.
>>>
>>> http://www.nabble.com/Re:-Re:-Re:-64-bit---double-question-p17329282.html
>>>
>>> Saludos,
>>> Felipe Sateler
>>>
>>
>>
>> --
>> Michael Gogins
>> Irreducible Productions
>> http://www.michael-gogins.com
>> Michael dot Gogins at gmail dot com
>>
>> ------------------------------------------------------------------------------
>> Register Now & Save for Velocity, the Web Performance & Operations
>> Conference from O'Reilly Media. Velocity features a full day of
>> expert-led, hands-on workshops and two days of sessions from industry
>> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
>
>
> --
>
>
> Andrés
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations
> Conference from O'Reilly Media. Velocity features a full day of
> expert-led, hands-on workshops and two days of sessions from industry
> leaders in dedicated Performance & Operations tracks. Use code vel09scf
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations
Conference from O'Reilly Media. Velocity features a full day of
expert-led, hands-on workshops and two days of sessions from industry
leaders in dedicated Performance & Operations tracks. Use code vel09scf
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-devel 



------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Yes, I also agree that there are too many options in the Csound build system.

Since the purpose of Csound is to include all possible audible and
musical bells and whistles, I believe that all features that are
available on all platforms should be mandatory in the build.

Also, I would prefer it if other features, that are not available on
all platforms, even if they are highly desirable and widely used on
some platforms, were required to be plugins.

What do you think?

Regards,
Mike

On 4/29/09, Felipe Sateler  wrote:
> I think it will depend mostly on the bus size than on the OS used. I would
> expect 32bit Linux, Windows and Mac to behave similarly.
>
> BTW, all this compile-time stuff add lots of complexity, which (IMO) is not
> necessary. Probably people will not like this, but I think it is much better
> if everything is required at compile-time, but optional at runtime. The
> complexity added by this approach is already solved by the plugin mechanism
> (plugins which can't be loaded just aren't). Of course, this makes it a bit
> more painful for people that build their own binaries, but if their need is
> not a highly specialized one, the distributed binary should be fixed instead
> of worked around. I would push for less options in the build stage. Of
> course, stuff like using doubles of floats cannot be chosen on runtime, but
> most other options can be eliminated by making them mandatory, which would
> not affect end-users.
>
> El 30/04/09 00:40 Michael Gogins escribió:
>> I propose to deprecate the float version, to distribute only the
>> double version, and to keep the float version around in case anyone
>> really needs it.
>>
>> My results are that the float version is somewhat faster on Windows, I
>> hear that the double version is quite a bit faster on Linux, and I
>> would like to know what the story is on the Mac.
>>
>> I do not wish to do anything hasty or that would alienate users, but I
>> think the current situation is not exactly optimal.
>>
>> Regards,
>> Mike
>>
>> On 4/29/09, Andres Cabrera  wrote:
>> > The floats build will still be of use for older machines. Will the
>> > floats option be completely removed, or the idea is to distribute only
>> > the doubles version, but have the floats version as an alternative
>> > compile option?
>> >
>> > Cheers,
>> > Andrés
>> >
>> > On Wed, Apr 29, 2009 at 7:19 AM, Michael Gogins
>> >
>> >  wrote:
>> >> Thank you. On Linux, then, the decision would seem not to require any
>> >> brains.
>> >>
>> >> I will perform similar tests myself on both Linux and Windows, when I
>> >> get the time, and obviously this kind of thing does not actually take
>> >> a lot of time to do.
>> >>
>> >> Regards,
>> >> Mike
>> >>
>> >> On 4/28/09, Felipe Sateler  wrote:
>> >>> El 29/04/09 06:02 Michael Gogins escribió:
>> >>>> I beg your indulgence as I repeat my call to omit the float version
>> >>>> of
>> >>>> csound, and to produce only the double version.
>> >>>>
>> >>>> Alternatively, to omit the double version, and produce only the float
>> >>>> version.
>> >>>>
>> >>>> The criteria would be, of course, to measure how fast each build
>> >>>> runs,
>> >>>> and do double-blind listening and mathematical comparison tests to
>> >>>> see
>> >>>> how good each build sounds.
>> >>>>
>> >>>> The current situation is a big waste of time unless, and only unless,
>> >>>> one of the builds is hopeless at efficiency, and the other is
>> >>>> hopeless
>> >>>> at precision.
>> >>>
>> >>> In 64bit linux, doubles seem to be even faster.
>> >>>
>> >>> http://www.nabble.com/Re:-Re:-Re:-64-bit---double-question-p17329282.ht
>> >>>ml
>> >>>
>> >>> Saludos,
>> >>> Felipe Sateler
>> >>
>> >> --
>> >> Michael Gogins
>> >> Irreducible Productions
>> >> http://www.michael-gogins.com
>> >> Michael dot Gogins at gmail dot com
>> >>
>> >> ------------------------------------------------------------------------
>> >>------ Register Now & Save for Velocity, the Web Performance &
>> >> Operations
>> >> Conference from O'Reilly Media. Velocity features a full day of
>> >> expert-led, hands-on workshops and two days of sessions from industry
>> >> leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> >> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> >> _______________________________________________
>> >> Csound-devel mailing list
>> >> Csound-devel@lists.sourceforge.net
>> >> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >
>> > --
>> >
>> >
>> > Andrés
>> >
>> > -------------------------------------------------------------------------
>> >----- Register Now & Save for Velocity, the Web Performance & Operations
>> > Conference from O'Reilly Media. Velocity features a full day of
>> > expert-led, hands-on workshops and two days of sessions from industry
>> > leaders in dedicated Performance & Operations tracks. Use code vel09scf
>> > and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
>> > _______________________________________________
>> > Csound-devel mailing list
>> > Csound-devel@lists.sourceforge.net
>> > https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
>
> Saludos,
> Felipe Sateler
>


-- 
Michael Gogins
Irreducible Productions
http://www.michael-gogins.com
Michael dot Gogins at gmail dot com

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net