Intel® C++ Compiler

Incomplete C++11 "alignas" implementation


icc 15.0 gives me a warning when compiling the following code:

using type alignas(64) = T;

warning #3463: alignas does not apply here
using type alignas(64) = T;

If I understand correctly my code is correct, and indeed GCC and Clang compile it properly without warnings. Is this a known issue with icc? Can I expect a fix in future versions?


Beginner choice

I'm learning WinAPI and like to play around with C++ for it.
Does the compiler really matters to this kind of task?
What Intel's compiler can do beyond MS C++? What is it for, exactly?
I also want to do some OpenGL programming? Does it help on it?

FATAL ERROR - ICC generates dysfunctional code on OSX 32-bit (plus ICC crashes when compiling debug version)

I use ICC on Windows and OSX. On Windows everything is fine. On OSX my project (release version) compiles just fine on 64-bit architecture. But on 32-bit it creates a malfunctioning code that crashes with a call stack containing a single line, saying that it crashed in operator delete(void*, void*). Note, it's the same code (except system services of course) on Windows 32 & 64 and Mac 32 & 64 and ONLY Mac 32 doesn't work. I'm having everything up to date (XCode, OSX, ICC), ICC is called from commandline.

Switching back to vc++: Cannot get rid of /fast option

I've been experimenting with the Intel compiler for the past few weeks, and decided to go back to the VC++ compiler.  I right clicked on the solution and selected "use vc++" compiler.  Tried to build, and got a bunch of warnings about "ignoring unknown option /fast".  I uninstalled the Intel compiler, completely blew away the development directly, got latest from source control, and the warnings are still there.  I searched all of the files in the dev directory (including property sheets and project settings) for the /dev flag.  Nothing. 

VADDSSL instruction?

Dear Intel developers,

I'm using intel 15 on E5-2670 processor. Analyzing my code by using Vtune, in a particolar line when I unpack a m128 type in order to sum in a single floating point each elements like horizontal sum, like this:


_mm_store_ps(denom_arr_tmp, denom_tmp);

 semblance[m_local] += denom_arr_tmp[0]+denom_arr_tmp[1]+denom_arr_tmp[2]+denom_arr_tmp[3];


The assembly generated is:

Initialized Lambda Capture

According to ICC15 should support initialized lambda captures. Yet code like this





*this = make([f, theLocalData = std::move(theLocalData)] {

fails to compile:

error: expected a "]"

1>        *this = make([f, theLocalData = std::move(theLocalData)] {

1>                                      ^


Intel® C++ Compiler abonnieren