Intel® Cilk™ Plus

avx512 mangling

Is there a more recent document than the Intel Vector ABI 2012-v0.9.5 document?

Has anyone decided on mangling for for avx512?  Right now there is:

'x' // xmm (SSE2) 
| 'y' // ymm1 (AVX1) 
| 'Y' // ymm2 (AVX2) 
| 'z' // zmm (mic)

Any mangling update?  ABI update?


intel vector abi mangling questions


I have a few questions regarding vector function mangling as specified in 2012-v0.9.5.  If this is not the right forum for this, could I please be guided to the appropriate one?

1. What is the order of the vparameters and the alignment parameter ('s', 'l', 'u', 'v', 'a')?  For example, if an argument is constantly linear by 2 and uniform, is that encoded as "l2u" or "ul2".  Similarly for the rest of the encoding characters.

vector/elementals attribute syntax on GCC

Hi folks.  Me again; trying to incorporate the elemental feature into GCC trunk.

The attribute syntax, as specified in the language spec is a bit problematic for the GCC infrastructure, and I'm wondering how committed are you to the current syntax.  For example, this is a sample vector use as allowed by the syntax:

__attribute__ ((vector(vectorlength(4), nomask, processor (core_i7_sse4_2), uniform (x,y))))
int foobar(int x, int y)

problem with parallelisation with Intel cilk plus,

Hey Guys,

I am working on a parallelisation of MJPEG Encoder using intel cilk plus framework. I have some data races problems while using "cilk_for" construct in this piece of code:

#include <stdio.h>

#include <stdlib.h>

#include <cilk/cilk.h>

TBlocks block; TPackets stream;

cilk_for (int t = 0; t < NumFrames; t++){

             for(int p = 0; p < VNumBlocks; p++) {

                    for( int k = 0; k < HNumBlocks; k++) {


NUMA-aware assignment of tasks to Cilk++ workers

I have a NUMA system. There is a thread for each core in the system. Threads that process similar data are assigned to the same node to reuse the data in the large L3 cache of the node. I want threads that are assigned to the same node should steal each other's jobs. If all jobs on a node have finished, these threads should steal jobs assigned to threads on other nodes. How can I implement this via Cilk++?

CilkPlus and SVN error

Hello everyone,

I was trying to update my SVN branch of cilkplus but I had an error with the server. The error was "could not connect to server (". I tried many times without success. Also, I tried to do this:

svn co cilkplus-gcc

with the same result.

Is there any other alternative to update the cilkplus repo?


Subscribe to Intel® Cilk™ Plus