topic Bug report: WinKaiser results contain NAN on x64 platform in Intel® Integrated Performance Primitives

Bug report: WinKaiser results contain NAN on x64 platform

devtree — Fri, 20 Aug 2010 22:23:56 GMT

Hi there.

I have a simple example below that exhibits this problem I'm having with the WinKaiser function. I get a NAN in the middle of my output array, with no error indication from IPP.

I tested it on multiple x64 processors. The code does notwork in x64, but does not exhibit the error on ia32. a particular I am using IPP 6.1 Build 137.56.

FYI...I checked the 6.1 errata list and did not see this error addressed. Does anybody know more about this issue? Thanks!

Code below:

void simpleWinKaiserTest()
{
// For len = 278650,the Kaiser window has 1.#QNaN values at ~len/2
// The IppStatus does not reflect an error condition.
// This occurs when using the IPP 6.1 build 137.56 (em64t),
//Tested withXeon E5520 and Celeron 550 processors
// Note: The same example works without error on ia32
/*
Actual returned output:

ippsSet_32f() IppStatus returned 0
ippsWinKaiser_32f_I() IppStatus returned 0
pDst[139324] = 1.#QNAN
pDst[139325] = 1.#QNAN
*/
IppStatus s;
Ipp32s len = 278650;
Ipp32f* pDst = ippsMalloc_32f(len);
if (pDst)
{
s = ippsSet_32f(1, pDst, len);
printf("ippsSet_32f() IppStatus returned %d\\n", s);

IppStatus s=ippsWinKaiser_32f_I(pDst, len, 0.0001f);
printf("ippsWinKaiser_32f_I() IppStatus returned %d\\n", s);

Ipp32s badIndex = 139324;
printf("pDst[%d] = %1.5f\\n",badIndex, pDst[badIndex]);
badIndex++;
printf("pDst[%d] = %1.5f\\n",badIndex, pDst[badIndex]);
ippsFree(pDst);
}
else
printf("malloc error\\n");
}

Bug report: WinKaiser results contain NAN on x64 platform

Gennady_F_Intel — Sat, 21 Aug 2010 06:36:56 GMT

Hi Soraya,

this is an unknown problem.

what is the linking line?

what IPP version you are using? ( please look in ..\doc\ippsupport.txt file and what isPackage ID?)

--Gennady

Bug report: WinKaiser results contain NAN on x64 platform

Gennady_F_Intel — Mon, 23 Aug 2010 05:34:31 GMT

Soraya,

I reproduce the problem with IPP 6.1 Upodate 5, dynamic linking with em64t libs on Core 2 Duo CPU. The same result happaned when out of place Kaiser function has been tested.

We will investiagate the cause if the probelm and back to you asap.

--Gennady

Bug report: WinKaiser results contain NAN on x64 platform

Gennady_F_Intel — Mon, 23 Aug 2010 11:48:54 GMT

Hi Soraya, just for your information.This issue has been submitted to our internal development tracking database for further investigation, we will inform you once a new update becomes available.

Here is a bug tracking number for your reference: DPD200191191.
Regards, Gennady

Bug report: WinKaiser results contain NAN on x64 platform

devtree — Tue, 24 Aug 2010 19:54:30 GMT

Hi Gennady,

Thanks for looking into it. I will wait from the outcome of the investigation.

Regards,
- Soraya

Bug report: WinKaiser results contain NAN on x64 platform

devtree — Wed, 10 Nov 2010 08:16:48 GMT

Hello there,

Is there any update on this bug?

Thanks and Regards,
- Soraya

Bug report: WinKaiser results contain NAN on x64 platform

Gennady_F_Intel — Wed, 10 Nov 2010 08:33:16 GMT

The fix of that problem will available in the version 7.0 Update 2.

Bug report: WinKaiser results contain NAN on x64 platform

Gennady_F_Intel — Thu, 10 Feb 2011 13:20:21 GMT

Hello Soraya,

IPP 7.0 Update 2 is available now,the issue should be fixed in the vesion. please refer to bleow information to get more information

Version 7.0.2 of the Intel Integrated Performance Primitives (Intel IPP) library is now available. The library is available as astand-alone productor as a component inIntel Parallel Studio 2011,Intel Parallel Studio XE 2011,Intel C++ Studio XE 2011,Intel Composer XE 2011, andIntel C++ Composer XE 2011. Please visit theIntel Software Evaluation Centerto download evaluation versions of these and other Intel software products.

What's New in 7.0.2

Additional optimizations for the256-bit Intel AVXSIMD instruction set.
FurtherIntel Atom Processor optimizationshave been incorporated.
Data compression gains of ~2x via a redesign of the "Inflate for Fixed Huffman" algorithm.
Cryptography improvements include 1.4x faster "sign verify" and a 2x faster ippsMontForm() function.
The ipp_zlib library is now based on the zlib 1.2.5 distribution.
Review a complete listof changes, bug fixes and new features on-line.

--Gennady