Intel® Integrated Performance Primitives

No threads being created when ippSetNumThreads is called

I have been trying to compress and decompress a text file using the LZO algorithm that IPP provides. It works fine when single threaded.

I tried using the function ippSetNumThreads() with arguments greater than 1.
Inorder to check if the specified number of threads are created I used the ippGetNumThreads() function.
The result is always 1 :( 

IPP UIC Performing very slow !!!

rudy B. 于 提交

Hi,

I am trying to observe the performance level of the IPP, for decoding purposes. I installed all the IPP package correctly, and was able to do some encodin/decoding tests with different file formats.

Any ways, what I am interested is to see how IPP performs for decdoing JP2 stream.

I have an ASUS G74S computer, which has 8 Intel cores: Intel(R) Cpre(TM) i7-2630QM CPU @ 2.00 GHz

The file under test is an encoded jp2 stream, and the file size is about 230KB (and just for sake of information, it is a1920x1080 size image).

Failure to decode (UIC Failure)

rudy B. 于 提交

Hi, 

I installed ipp samples (7.1.1) correctly. And, I used UIC examples.

It works fine for only certain images. To be exact, it only worked for the images encoded using UIC.

For example, I had few "jpeg" images, and then using the IPP UIC tool, I enocded them to "jp2" format. Then I used the generated "jp2" format to convert it to different formst (e.g. "bmp", "png" or even back to "jpeg"). And they all worked fine.

Regarding the performance of function ippsDecodeLZO_8u()

From my understanding IPP libraries use SIMD instructions so that they outperform other libraries. So I tried using the 'disass' feature of gdb to check if there are any SIMD instructions present in the assembly code of the function ippsDecodeLZO_8u().

0x00000000004010a0 <ippsDecodeLZO_8u+0>:        mov    0x207ab1(%rip),%rax        

0x00000000004010a7 <ippsDecodeLZO_8u+7>:        mov    (%rax),%eax

0x00000000004010a9 <ippsDecodeLZO_8u+9>:        mov    0x207a98(%rip),%r11       

Max / Min of two vectors

a b. 于 提交

Hello everyone,

I want to know if Intel IPP offers a function to compute the componentwise maximum (minimum) of two vectors, i.e.

c=max(a,b) <--> c_i =max(a_i,b_i)

where a and b are vectors of float (32f) or doubles (64f).

I could not find it in the reference manual.

Sincerely,

M. Ampf

Can't build audio-video-codecs folder

Hello,

I followed the different steps to setup the main sample (ipp-samples.7.1.1.013) and I'm trying to build the audio-video-codecs by typing the line "perl build.pl --cmake=audio-video-codecs,ia32,vc2008,s,st,debug"

Build fails with the following error:

"6>..\..\..\..\sources\audio-video-codecs\application\common\src\umc_app_utils.cpp(44) : error C2065: 'ippCpuAVX2' : undeclared identifier"

I am using all the software required (Intel® Composer XE 2011 SP1, CMake* 2.8.8, Perl 5.12.2), except IPP 7.0 instead of IPP 7.1. Could that explain the error?

页面

订阅 Intel® Integrated Performance Primitives