Quantcast
Channel: Intel® Integrated Performance Primitives
Viewing all 1294 articles
Browse latest View live
↧

MIC - Many Integrated Core Offloading

December 4, 2014, 8:52 am
$
0
0

I am haveing trouble compiling a program that offloads ipp function to the mic. I just need something simple like this:

What linking and compiling options need to be enabled?

Currently it just says:

error: undefined reference to `ippsMul_32f'

<pre class="brush:cpp">

double mult_mic_ipp(float* a,float* b, float *c, int nsize)

{

    double start = 0;
    double stop = 0;
    int numt = 0;
 
#pragma offload target (mic) in(a,b:length(nsize)) out(c:length(nsize)) inout(start,stop,numt)
{
    start = omp_get_wtime();
    ippsMul_32f(a,b,c,nsize);
    stop = omp_get_wtime();
}
    return stop-start;
}

</pre>

 

↧
Search

link error: libipps.a, IPP 8.2

December 8, 2014, 8:16 am
$
0
0

After migrating to Mac OSX 10.10,  because composer_xe_2013 doesn't support osx 10.10. We also  update intel compiler to composer_xe_2015.

But when link  libipps.a with our project using clang+ and osx sdk 10.10,  we met this link error;  Could you take a look what problem it might be? since it works with  previous intel compiler.

Undefined symbols for architecture i386:
  "_g9_ippsCos_32f_A24", referenced from:
      _g9_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_g9---ipps_initTabTwdBase_32f.o)
  "_g9_ippsSin_32f_A24", referenced from:
      _g9_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_g9---ipps_initTabTwdBase_32f.o)
  "_p8_ippsCos_32f_A24", referenced from:
      _p8_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_p8---ipps_initTabTwdBase_32f.o)
  "_p8_ippsSin_32f_A24", referenced from:
      _p8_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_p8---ipps_initTabTwdBase_32f.o)
  "_px_ippsCos_32f_A24", referenced from:
      _px_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_px---ipps_initTabTwdBase_32f.o)
  "_px_ippsSin_32f_A24", referenced from:
      _px_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_px---ipps_initTabTwdBase_32f.o)
  "_v8_ippsCos_32f_A24", referenced from:
      _v8_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_v8---ipps_initTabTwdBase_32f.o)
  "_v8_ippsSin_32f_A24", referenced from:
      _v8_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_v8---ipps_initTabTwdBase_32f.o)
  "_w7_ippsCos_32f_A24", referenced from:
      _w7_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_w7---ipps_initTabTwdBase_32f.o)
  "_w7_ippsSin_32f_A24", referenced from:
      _w7_ipps_initTabTwdBase_32f in libipps.a(psfftsblkcn_w7---ipps_initTabTwdBase_32f.o)
ld: symbol(s) not found for architecture i386

my IPP version info:

Package ID:  m_ipp_8.2.1.108 m_ccompxe_2015.1.108

This package contains:
    Intel(R) Integrated Performance Primitives 8.2 Update 1 for OS X*

 

 

 

 

 

 

↧
↧

MP3 & AAC decoding, and advantages of IPP 8

December 13, 2014, 4:55 am
$
0
0

Hi, 

I own IPP 6 for Windows & Mac and use it for signal processing extensively. Now I need to provide MP3 and AAC decoding (potentially other formats, but these are the important ones). I found there is "UMC" - is this the solution to look for? Does it work on both Windows and Mac, both 32 bit and 64bit? It's not a problem to upgrade to IPP 8 if needed (that's the newest version). 

Why isn't it included in IPP itself and rather part of some kind of examples? How to compile it effectively? It looks like lots of separate projects.

Which comes to one more question, a little off-topic - are there performance improvements for IPP 8 compared to IPP 6? Mainly about 1D basic processing (mul, add, complex mul, IIR filters, FFT). 

Thanks in advance!

↧

Media SDK crossplatform = for windows???

December 17, 2014, 9:46 am
$
0
0

Hi,

I need an mp3 decoder, for windows and mac os x, I own IPP 6, but it seems that UMC has been deprecated (??) and is not available for OSX 64-bit, so the alternative is "Media SDK", but:

https://software.intel.com/en-us/media-client-solutions

There is:

"Media SDK for Windows* - A cross-platform API for developing media apps (foundation)." - What on earth is that? Is it crossplatform then or is it not? Is it available for OSX 32 & 64?

↧

ippsHMACSHA1MessageDigest fails with ippStsLengthErr

December 22, 2014, 7:54 am
$
0
0

Hi,

I'm using the function ippsHMACSHA1MessageDigest(const Ipp8u *pSrcMesg, int mesgLen, const Ipp8u *pKey, int keyLen, Ipp8u *pMAC, int macLen) and it fails with ippStsLengthErr.

The parameters are : mesgLen = 176, keyLen = 20, macLen = 10.

What could be the reason for ippStsLengthErr?

Thanks for the help!

↧
↧

2D Convolution performance between 8Bit and 16Bit

December 27, 2014, 1:06 am
$
0
0

Hi,

For IPP version 8.1 or 8.2, I would like to find out if there should be any big differences in speed performance for  2D Convolution function between Unsigned 8Bit and Signed 16Bit fixed point data? Should the Unsigned 8Bit Convolution function be 2 times or even 4 timer faster  than Signed 16Bit Convolution function?

Thanks!

↧

ippsSortRadixIndexAscend 64-bit

December 28, 2014, 3:49 am
$
0
0

Hi guys,

Is there a 64-bit version of ippsSortRadixIndexAscend available, or otherwise are there any alternatives available to me for 64-bit radix sorting, either in the IPP or MKL?  Right now I need to switch down to 32-bit, which is losing some accuracy in my sorting.

Warm regards,

Mark

↧

segfault in ippiResizeSqrPixel

December 30, 2014, 6:39 pm
$
0
0

hi. i'm using ipp6(6.1.2) for image processing to make thumbnail.
and my env is that.

OS : linux(centOS 5.X) 64bit
CPU : Intel(R) Xeon(R) CPU E5-2630L 0 @ 2.00GHz

And I got a trouble during upgrage  centOS 5.x --> 6.x

there is no segmentation fault to execute ippiResizeSqrPixel_8u_C3R in CentOS5.X.
but, here is a segmentation fault in CentOS 6.X

of cause, I re-install IPP libs, and re-compile my binary.

 

here is a segfault stack.

#0  0x0000000000502e19 in e9_ownpi_CalcBorder8pxLz ()
No symbol table info available.
#1  0x00000000004ebf04 in e9_ownpiResizeSP ()
No symbol table info available.
#2  0x00000000004c75c9 in e9_ippiResizeSqrPixel_8u_C3R ()

 

What's unique is segfault occur in only special conditions.
============================================
Origial Image Width : 15
Origial Image Hieght : 15
Target Resize Width : 176
Target Resize Height : 176
============================================

SUMMERY

1. when I change target size, it doesn't make segfault.
2. when I execute it in centOS 5.X, it doesn't make segfault.
3. when I change CPU(with CentOS 6.X), it  doesn't make segfault.
 

I guess CPU and IPP and OS has some dependency.
Would somebody help me please?

 

 

↧
Search

Where could I get Intel Integrated Performance Primitives v4.0?

December 31, 2014, 12:01 am
$
0
0

Running a software need this lib, but I cannot find it both in intel sites and search at google. Can anyone help?

↧
↧

Finding the largest matching rectangle between two 8-bit images

January 1, 2015, 5:51 am
$
0
0

Hi,

Say there are two images A and B, where B is part of A centered around a matching pixel P. B is not necessarily entirely contained within A, but is guaranteed to contain part of A.

I would like to find the largest matching rectangle of B within A. I have implemented various non-IPP algorithms that does this in a relatively efficient manner, but would like to see whether IPP can help in anyway.

I've looked at the IPP function list and nothing seems to do what I want, which is effectively to find an ROI (most take an ROI as input!). ippiCompare_8u returns an 8-bit bitmap with 0xFF for a matching pixel, and 0 for a mismatched pixel. Finding the largest rectangle is then possible, however, a suitable maximal rectangle algorithm needs to be used which is time consuming.

Is there a combination of IPP functions that will give me what I'm looking for?

Thanks in advance!

↧

IPP OpenCL and Asynchronous C++ drivers

January 1, 2015, 9:50 am
$
0
0

Hello,

I have Intel Composer XE 2015 which contains IPP 8.2.  I am unable to locate any of the OpenCL or asynchronous C++ code.  Is this code included with IPP 8.2.  If not, where can I get it?

In general, how does intel recommend using IPP with GPU support if it exists.  Moving forward would it be better to use OpenCL?

Regards,

Brad Kimbrough

↧

ippAlgDirect and ippAlgFFT take the same time

January 8, 2015, 4:30 am
$
0
0

Hello,

I'm running  ippiCrossCorNorm_16u32f and  ippiCrossCorNorm_32f  on ipp 8.1

It seems that timing is related only to src and template sizes.

I tried running with ippAlgDirect and ippAlgFFT but timing is  the same in both methods.

timing is also the same for 16u input and float input

Im I doing something wrong?

 

Thanks

 

↧

Intel IPP G723 and G729 audio codecs license

January 8, 2015, 10:02 am
$
0
0

Hi

I want to integrate Intel IPP G723 and G729 audio codecs into pjsua. This will be used into a commercial application that needs to be installed on multiple places. We would like to know:

- how much a license for  IPP G723 and G729 audio codecs costs

- what types of licenses are available for multiple installations and how much it costs in each case

thanks

Bogdan

↧
↧

What is the ImageDataPtr class

January 10, 2015, 2:58 am
$
0
0

Hello,

I need to write code for the JPEG encryption. I got the sample source code. To write my client app, I need to provide the parameters. B

I am struct-king with providing one parameter i,.e ImageDataPtr class method i,.e

 

UIC::ImageDataPtr dataPtr;
    dataPtr.p8u = //what I need to provide here;

could you please suggest me.

 

Regards,

sathish.

 

↧

how to use IPP with opencv

January 11, 2015, 10:44 pm
$
0
0

hi guys,

i have intel cluster studio 2013 and redhat OS with 2 computenode having one intel mic card and one gpu (NVIDIA k20x). i tried so many ways to install opencv in masternode but i enable to proceed please do you have any idea about how to install opencv using intel parallel complier

please help me 

↧
Search

IppiRotate IppiGetRotateShift

January 13, 2015, 4:50 am
$
0
0

Hi

From Ipp 8.2 documentation

1) IppiRotate :

angle : The angle of rotation in degrees. The source image is rotated counterclockwise around the origin (0,0).

IppiRotateCenter : ... This function rotates the ROI of the source image by angle degrees (counterclockwise for positive angle values)

while IppiGetRotateShift and AddRotateShift:

angle : The angle in degrees to rotate the image clockwise around the point with coordinates (xCenter, yCenter).

Positive angle is in opposite direction for the last 2 function?

2) While not documented function IppiRotateCenter can return also ( as warning I suppose) ippStsWrongIntersectROI. Is right?

Thank's

Giacomo C.

 

 

 

↧

Unresolved external ippmInvert_m_64f@32

January 13, 2015, 5:28 am
$
0
0

Hi, I switched (or am trying to) to the new ipp version coming with the coposer 2015. so ipp8.2. I get an unresolved exteran on the ippm invert function. Non of . .lib seems to contain it. I know it is deprecated but it should still be there...

So where has it gone?

Any help would be appreciated.

Regards,

Rob

 

 

 

 

↧
↧

standalone version confusion & request

January 14, 2015, 1:19 pm
$
0
0

Hi,

https://software.intel.com/en-us/intel-ipp states "Now available as a standalone product for OS X". But https://software.intel.com/en-us/articles/intel-sdp-changes#IPSXEingredi... states "New licenses of the following ingredient products will be available only in suites going forward.". Which of the two is true?

I think, it would be really unfortunate, if the standalone versions would be no longer available. I'd love to use IPP for power of two 1d FFT anf iFFT on windows. On OSX I can use apples accelerate framework. But on windows there's to my knowledge nothing comparable. But I'd also consider using it on OSX because of the support for non power of two FFTs.

Assuming, that the windows version is no longer available as a standalone product. What is the most cost effective option right now to license IPP on Windows? I'm working on a tight budget and USD 500 or more is a little bit on the steep side of things.

Thanks for the clarification and suggestions
-- Benjamin

 

↧

ippiBGRToLab_32f_C3R Segfault on non-scaled data

January 14, 2015, 2:49 pm
$
0
0

The BGR to Lab documentation mentions that

 The equations above are given on the assumption that initial B, G, R values are normalized to the range [0..1]. The computed values of the L component are in the range [0..100], a and b component values are in the range [-128..127].

Is this intended to mean that the conversion functions cannot be used if the floating point data does not lie in that range? Either way, these calls on my system using the edition of IPP that shipped with composer_xe_2015.1.133 cause a segmentation fault if the floating point data does not lie in that range.

↧

ippiGaussianBorder now only supports const borders

January 16, 2015, 2:42 am
$
0
0

Hi,

I'm switching to the new Composer 2015 package containing IPP 8.2. I ran into a 'bug' straight away. A crucial part of my application suddenly didn't work anymore. Cause: see the title. I use mirrored borders...:-(. I cannot find any information on this change, and I am also worried that there will be more of this kind of surprises. Is there a list of functions of which the functionality has been limited?

Btw. I am worried about the future of IPP. I seems Intel is planning to deprecate the about the whole product. I also think that Intel should supply examples of how to code alternatives for deprecated functions in the documentation.

Regards,

Rob

 

 

↧
Viewing all 1294 articles
Browse latest View live
Search