Xojo Developer Conference
25/27th April 2018 in Denver.
MBS Xojo Conference
6/7th September 2018 in Munich, Germany.

[MBS] LCMS2 Performance (MBS Xojo Plugin Mailinglist archive)

Back to the thread list
Previous thread: [MBS] OverlayMBS on Windows: Shadow?
Next thread: [MBS] Error generating PDF in web edition


Re: [MBS] PortAudio bug?   -   Garth Hjelte
  [MBS] LCMS2 Performance   -   Lee Badham
   Re: [MBS] LCMS2 Performance   -   Christian Schmitz
   Re: [MBS] LCMS2 Performance   -   Lee Badham
   Re: [MBS] LCMS2 Performance   -   Christian Schmitz
   Re: [MBS] LCMS2 Performance   -   Lee Badham
   Re: [MBS] LCMS2 Performance   -   Lee Badham
   Re: [MBS] LCMS2 Performance   -   Christian Schmitz
   Re: [MBS] LCMS2 Performance   -   Lee Badham
   Re: [MBS] LCMS2 Performance   -   Lee Badham

[MBS] LCMS2 Performance
Date: 23.01.13 18:08 (Wed, 23 Jan 2013 17:08:54 +0000)
From: Lee Badham
Hi,

I've been updating an app to use LCMS2, rather than LCMS1 and have a 2000% performance drop using CreateTransform.

Basically:

Function Createlcms2transform(p as LCMS2ProfileMBS) As LCMS2TransformMBS
return LCMS2TransformMBS.CreateTransform(p,LCMS2MBS.kTYPE_CMYK_16,lcms2labprofileLCMS2MBS.kTYPE_Lab_16,3)

End Function

Takes 20 times longer than

Function Createlcmstransform1(p as CMProfileMBS) As CMTransformMBS
return LCMSMBS.CreateTransform(p,LCMSMBS.TYPE_CMYK_16,lcmslabprofile,LCMSMBS.TYPE_Lab_16,3)

End Function

or

Function Createlcmstransform(p as CMProfileMBS) As CMTransformMBS
return LCMSMBS.CreateTransform(p,lcmslabprofile,3)

End Function

Here's the profile information calling these functions 100 times.

Window1
Name Called Average (seconds) Total (seconds) %
Createlcmstransform 101 0.0042 0.42 3%
Createlcmstransform1 101 0.004 0.40 3%
Createlcms2transform 101 0.1064 10.75 90%
Total 23.15 193%

What can I do about this? As I import colour information I create these transforms to calculate new colours so each transform is done many times.

Lee Badham

www.bodoni.co.uk | www.presssign.com

_______________________________________________
Mbsplugins_monkeybreadsoftware.info mailing list
<email address removed>
https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Re: [MBS] LCMS2 Performance
Date: 28.01.13 16:39 (Mon, 28 Jan 2013 16:39:46 +0100)
From: Christian Schmitz

Am 28.01.2013 um 13:36 schrieb Lee Badham <<email address removed>>:

> Hi,
>
> Here's the link:
>
> It shows a listbox with 3 calculations of the Lab value, and also the memory block values.

Thanks. Will check this.

Greetings
Christian

Re: [MBS] LCMS2 Performance
Date: 28.01.13 13:36 (Mon, 28 Jan 2013 12:36:51 +0000)
From: Lee Badham
Hi,

Here's the link:

http://www.presssign.com/downloads/presssign/lcms1_2.rbp.zip

It shows a listbox with 3 calculations of the Lab value, and also the memory block values.

Lee

On 28 Jan 2013, at 12:25, Christian Schmitz <<email address removed>> wrote:

>
> Am 28.01.2013 um 13:24 schrieb Lee Badham <<email address removed>>:
>
>> Annnndddd…….
>>
>> The numbers are different :(
>>
>> Converting a CMYK value to Lab value gives different numbers between LCMS1 and LCMS2 with the same transformation parameters, whether I use cmsFLAGS_NOOPTIMIZE or not.
>>
>> Anyone else found any difference?
>
> Can you provide sample project?
>
> Greetings
> Christian
>
> --
> Read our blog about news on our plugins:
>
> http://www.mbsplugins.de/
>
> _______________________________________________
> Mbsplugins_monkeybreadsoftware.info mailing list
> <email address removed>
> https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

_______________________________________________
Mbsplugins_monkeybreadsoftware.info mailing list
<email address removed>
https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Re: [MBS] LCMS2 Performance
Date: 28.01.13 13:25 (Mon, 28 Jan 2013 13:25:19 +0100)
From: Christian Schmitz

Am 28.01.2013 um 13:24 schrieb Lee Badham <<email address removed>>:

> Annnndddd…….
>
> The numbers are different :(
>
> Converting a CMYK value to Lab value gives different numbers between LCMS1 and LCMS2 with the same transformation parameters, whether I use cmsFLAGS_NOOPTIMIZE or not.
>
> Anyone else found any difference?

Can you provide sample project?

Greetings
Christian

Re: [MBS] LCMS2 Performance
Date: 28.01.13 13:24 (Mon, 28 Jan 2013 12:24:41 +0000)
From: Lee Badham
Annnndddd…….

The numbers are different :(

Converting a CMYK value to Lab value gives different numbers between LCMS1 and LCMS2 with the same transformation parameters, whether I use cmsFLAGS_NOOPTIMIZE or not.

Anyone else found any difference?

Lee

On 24 Jan 2013, at 10:14, Lee Badham <<email address removed>> wrote:

> Hi,
>
> So adding cmsFLAGS_NOOPTIMIZE to the transform makes it faster, so we are back into LCMS1 performance. Now I've got to check whether the calculations are good / fast enough
>
> Lee
>
>
> On 23 Jan 2013, at 17:08, Lee Badham wrote:
>
>> Hi,
>>
>> I've been updating an app to use LCMS2, rather than LCMS1 and have a 2000% performance drop using CreateTransform.
>>
>> Basically:
>>
>> Function Createlcms2transform(p as LCMS2ProfileMBS) As LCMS2TransformMBS
>> return LCMS2TransformMBS.CreateTransform(p,LCMS2MBS.kTYPE_CMYK_16,lcms2labprofileLCMS2MBS.kTYPE_Lab_16,3)
>>
>> End Function
>>
>> Takes 20 times longer than
>>
>> Function Createlcmstransform1(p as CMProfileMBS) As CMTransformMBS
>> return LCMSMBS.CreateTransform(p,LCMSMBS.TYPE_CMYK_16,lcmslabprofile,LCMSMBS.TYPE_Lab_16,3)
>>
>> End Function
>>
>> or
>>
>> Function Createlcmstransform(p as CMProfileMBS) As CMTransformMBS
>> return LCMSMBS.CreateTransform(p,lcmslabprofile,3)
>>
>> End Function
>>
>> Here's the profile information calling these functions 100 times.
>>
>> Window1
>> Name Called Average (seconds) Total (seconds) %
>> Createlcmstransform 101 0.0042 0.42 3%
>> Createlcmstransform1 101 0.004 0.40 3%
>> Createlcms2transform 101 0.1064 10.75 90%
>> Total 23.15 193%
>>
>> What can I do about this? As I import colour information I create these transforms to calculate new colours so each transform is done many times.
>>
>> Lee Badham
>>
>> www.bodoni.co.uk | www.presssign.com
>>
>>
>> _______________________________________________
>> Mbsplugins_monkeybreadsoftware.info mailing list
>> <email address removed>
>> https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info
>
> Lee Badham
>
>
>
>
> t: 01923 220 530
> <email address removed>
> www.bodoni.co.uk | www.presssign.com
> skype: bodonilee
> Bodoni Systems Ltd
> Unit 19
> Orbital 25 Business Park
> Dwight Road
> Watford
> WD18 9DA
>
> _______________________________________________
> Mbsplugins_monkeybreadsoftware.info mailing list
> <email address removed>
> https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

_______________________________________________
Mbsplugins_monkeybreadsoftware.info mailing list
<email address removed>
https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Re: [MBS] LCMS2 Performance
Date: 24.01.13 11:19 (Thu, 24 Jan 2013 10:19:11 +0000)
From: Lee Badham
Hi,

So adding cmsFLAGS_NOOPTIMIZE to the transform makes it faster, so we are back into LCMS1 performance. Now I've got to check whether the calculations are good / fast enough

Lee

On 24 Jan 2013, at 10:10, Lee Badham wrote:

> Hi,
>
> Same performance hit on MBS 12.5 and MBS 13.1pr8, so LCMS 2.4 still has this problem.
>
> Lee
>
> On 23 Jan 2013, at 17:08, Lee Badham wrote:
>
>> Hi,
>>
>> I've been updating an app to use LCMS2, rather than LCMS1 and have a 2000% performance drop using CreateTransform.
>>
>> Basically:
>>
>> Function Createlcms2transform(p as LCMS2ProfileMBS) As LCMS2TransformMBS
>> return LCMS2TransformMBS.CreateTransform(p,LCMS2MBS.kTYPE_CMYK_16,lcms2labprofileLCMS2MBS.kTYPE_Lab_16,3)
>>
>> End Function
>>
>> Takes 20 times longer than
>>
>> Function Createlcmstransform1(p as CMProfileMBS) As CMTransformMBS
>> return LCMSMBS.CreateTransform(p,LCMSMBS.TYPE_CMYK_16,lcmslabprofile,LCMSMBS.TYPE_Lab_16,3)
>>
>> End Function
>>
>> or
>>
>> Function Createlcmstransform(p as CMProfileMBS) As CMTransformMBS
>> return LCMSMBS.CreateTransform(p,lcmslabprofile,3)
>>
>> End Function
>>
>> Here's the profile information calling these functions 100 times.
>>
>> Window1
>> Name Called Average (seconds) Total (seconds) %
>> Createlcmstransform 101 0.0042 0.42 3%
>> Createlcmstransform1 101 0.004 0.40 3%
>> Createlcms2transform 101 0.1064 10.75 90%
>> Total 23.15 193%
>>
>> What can I do about this? As I import colour information I create these transforms to calculate new colours so each transform is done many times.
>>
>> Lee Badham
>>
>> www.bodoni.co.uk | www.presssign.com
>>
>> _______________________________________________
>> Mbsplugins_monkeybreadsoftware.info mailing list
>> <email address removed>
>> https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info
>
> Lee Badham
>
> www.bodoni.co.uk | www.presssign.com
>
> _______________________________________________
> Mbsplugins_monkeybreadsoftware.info mailing list
> <email address removed>
> https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Lee Badham

www.bodoni.co.uk | www.presssign.com

_______________________________________________
Mbsplugins_monkeybreadsoftware.info mailing list
<email address removed>
https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Re: [MBS] LCMS2 Performance
Date: 24.01.13 11:18 (Thu, 24 Jan 2013 11:18:29 +0100)
From: Christian Schmitz

Am 24.01.2013 um 11:14 schrieb Lee Badham <<email address removed>>:

> Hi,
>
> So adding cmsFLAGS_NOOPTIMIZE to the transform makes it faster, so we are back into LCMS1 performance. Now I've got to check whether the calculations are good / fast enough
>

Oh, yes.

Prebuilding some lookup tables takes time of course.

Greetings
Christian

Re: [MBS] LCMS2 Performance
Date: 24.01.13 11:14 (Thu, 24 Jan 2013 10:14:23 +0000)
From: Lee Badham
Hi,

So adding cmsFLAGS_NOOPTIMIZE to the transform makes it faster, so we are back into LCMS1 performance. Now I've got to check whether the calculations are good / fast enough

Lee

On 23 Jan 2013, at 17:08, Lee Badham wrote:

> Hi,
>
> I've been updating an app to use LCMS2, rather than LCMS1 and have a 2000% performance drop using CreateTransform.
>
> Basically:
>
> Function Createlcms2transform(p as LCMS2ProfileMBS) As LCMS2TransformMBS
> return LCMS2TransformMBS.CreateTransform(p,LCMS2MBS.kTYPE_CMYK_16,lcms2labprofileLCMS2MBS.kTYPE_Lab_16,3)
>
> End Function
>
> Takes 20 times longer than
>
> Function Createlcmstransform1(p as CMProfileMBS) As CMTransformMBS
> return LCMSMBS.CreateTransform(p,LCMSMBS.TYPE_CMYK_16,lcmslabprofile,LCMSMBS.TYPE_Lab_16,3)
>
> End Function
>
> or
>
> Function Createlcmstransform(p as CMProfileMBS) As CMTransformMBS
> return LCMSMBS.CreateTransform(p,lcmslabprofile,3)
>
> End Function
>
> Here's the profile information calling these functions 100 times.
>
> Window1
> Name Called Average (seconds) Total (seconds) %
> Createlcmstransform 101 0.0042 0.42 3%
> Createlcmstransform1 101 0.004 0.40 3%
> Createlcms2transform 101 0.1064 10.75 90%
> Total 23.15 193%
>
> What can I do about this? As I import colour information I create these transforms to calculate new colours so each transform is done many times.
>
> Lee Badham
>
> www.bodoni.co.uk | www.presssign.com
>
> _______________________________________________
> Mbsplugins_monkeybreadsoftware.info mailing list
> <email address removed>
> https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Lee Badham



t: 01923 220 530
<email address removed>
www.bodoni.co.uk | www.presssign.com
skype: bodonilee
Bodoni Systems Ltd
Unit 19
Orbital 25 Business Park
Dwight Road
Watford
WD18 9DA

_______________________________________________
Mbsplugins_monkeybreadsoftware.info mailing list
<email address removed>
https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Re: [MBS] LCMS2 Performance
Date: 24.01.13 11:10 (Thu, 24 Jan 2013 10:10:16 +0000)
From: Lee Badham
Hi,

Same performance hit on MBS 12.5 and MBS 13.1pr8, so LCMS 2.4 still has this problem.

Lee

On 23 Jan 2013, at 17:08, Lee Badham wrote:

> Hi,
>
> I've been updating an app to use LCMS2, rather than LCMS1 and have a 2000% performance drop using CreateTransform.
>
> Basically:
>
> Function Createlcms2transform(p as LCMS2ProfileMBS) As LCMS2TransformMBS
> return LCMS2TransformMBS.CreateTransform(p,LCMS2MBS.kTYPE_CMYK_16,lcms2labprofileLCMS2MBS.kTYPE_Lab_16,3)
>
> End Function
>
> Takes 20 times longer than
>
> Function Createlcmstransform1(p as CMProfileMBS) As CMTransformMBS
> return LCMSMBS.CreateTransform(p,LCMSMBS.TYPE_CMYK_16,lcmslabprofile,LCMSMBS.TYPE_Lab_16,3)
>
> End Function
>
> or
>
> Function Createlcmstransform(p as CMProfileMBS) As CMTransformMBS
> return LCMSMBS.CreateTransform(p,lcmslabprofile,3)
>
> End Function
>
> Here's the profile information calling these functions 100 times.
>
> Window1
> Name Called Average (seconds) Total (seconds) %
> Createlcmstransform 101 0.0042 0.42 3%
> Createlcmstransform1 101 0.004 0.40 3%
> Createlcms2transform 101 0.1064 10.75 90%
> Total 23.15 193%
>
> What can I do about this? As I import colour information I create these transforms to calculate new colours so each transform is done many times.
>
> Lee Badham
>
> www.bodoni.co.uk | www.presssign.com
>
> _______________________________________________
> Mbsplugins_monkeybreadsoftware.info mailing list
> <email address removed>
> https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info

Lee Badham

www.bodoni.co.uk | www.presssign.com

_______________________________________________
Mbsplugins_monkeybreadsoftware.info mailing list
<email address removed>
https://ml01.ispgateway.de/mailman/listinfo/mbsplugins_monkeybreadsoftware.info