Intel® C++ Compiler

AVX2 _mm256_cmpeq_epi8/16/32/64 documentation specifies incorrect value in case of equality

In the User and Reference Guide for the Intel C++ Compiler 15.0, the description of the AVX2 intrinsics _mm256_cmpeq_epi8/16/32/64 states:

If a pair of data elements is equal, the corresponding data element in the destination vector is set to 1.

ICC 15: gfx_linker exits because binutils doesn't recognize .obj file. Is it possible to disable IGT code generation?

Hello,

When I'm attempting to compile Firefox with Intel C++ 15 Update 1 and Visual Studio 2010 I fall into the following error:

 1:01.09 ../../widget/windows/nsWindowGFX.obj: file not recognized: File format not recognized
 1:01.09 gfx_linker: catastrophic error: command 'ld.exe' exited with non-zero exit code 1

This seems to occur because code generation for Intel Graphics Technology (IGT) seems to use binutils and that doesn't seem to support windows object (.obj) files.

Suppressing vectorization remarks

Intel C++ 15.0 (Update 1) is spitting out many vectorization remarks of this form:

remark: loops in this subroutine are not good vectorization candidates (try compiling with O3 and/or IPO)

when I do builds for profiling with these options: 

icl /nologo /Qstd=c++11 /Qcxx-features /Wall /Wp64 /Qdiag-disable:177,869,1786,2259,3280,10382,11074,11075 /QxHOST /DNOMINMAX /DWIN3
2_LEAN_AND_MEAN /DNDEBUG /Qansi-alias /O3 /fp:fast /Qprec-div- /Qip /Z7 /c /object:myfile.obj myfile.cc

There are a number of problems with this:

AVX2 permute intrinsics and copying a source-vector element to multiple destination-vector elements

In the User and Reference Guide for the Intel C++ Compiler 15.0, the descriptions of the AVX2 intrinsics _mm256_permutevar8x32_epi32 and _mm256_permutevar8x32_ps state:

The intrinsic does NOT allow to copy the same element of the source vector to more than one element of the destination vector.

Purchase link of Intel® Parallel Studio XE 2013 Composer Edition for C++, OS X

Hi,

Where can I find the purchase link of Intel® Parallel Studio XE 2013 Composer Edition for C++, OS X? And after ordering edition 2013, if I want to upgrade to 2015, what's the price? Thx.

My mail: jackwl2010@163.com

Pure Virtual Function Incorrectly Inlined by IPO w/ Anonymous Union

It appears that the compiler is incorrectly inlining calls to pure virtual functions in certain situations where IPO is enabled and an anonymous union is declared. The simplest example I can come up with is below. Removing the anonymous union or compiling without IPO resolves the issue.This issue is present in compiler versions 14.0.1 and 15.0.1 on CentOS 6.6 with gcc 4.4. Please advise on any known workarounds or fixes in upcoming releases.

AVX2 Intrinsics for Arithmetic Shift Operations documentation discrepancy

In the User and Reference Guide for the Intel C++ Compiler 15.0, each of the AVX2 Intrinsics for Arithmetic Shift Operations describe the behavior of the "empty" bits resulting from an arithmetic right shift operation in a manner similar to the following (e.g., the description for _mm256_srai_epi16/32):

Intel® C++ Compiler abonnieren