Intel® Cilk™ Plus

skip cilk-plus in gcc testsuite?

Given that gcc cilk-plus isn't supported on Windows Subsystem for Linux, and plans have been advertised for its removal from gcc 8, is there a way to avoid running the failing cilk-plus tests in gcc testsuite?  Configuring --disable-libcilkrts stops the attempt to build libcilkrts but doesn't skip the failing tests.  I hadn't yet removed all copies of .a and .spec from the WSL installation

Regression on ARM hardware gcc 6.2 and 7.1

Thank you for including the slight modifications that allow gcc with Cilk support to build for ARM architectures.  I mentioned a slight problem with performance at and also at relating to the 7.1 version of the gcc Cilk.

Code Exploring in Cilkplus


CilkPlus comes with some keywords like cilk_spawn, cilk_for etc. I wonder if I can explore how are they invoked and where the definition for them are actually written. For example, when cilk_spawn is invoked where the control goes?

Beside this, inside cilk.h there are some macros defined as the same name as those keywords. But I don't know where the code for these things is actually written. I am facing difficulties to get the flow. Any idea regarding this? Please let me know if my question is not clear?




Does Intel Cilk Plus SDK support offload to gfx?


Environment: Linux, latest GCC.

I am evaluating the offloading capabilities of the Intel Cilk Plus SDK using: #pragma offload target(gfx)

Is it supported using the latest Cilk Plus SDK or is Intel Compiler must be used?

When trying the latest GCC it gives a warning that this pragma is not supported.


Tomer Gal

Using SVM for Intel Graphics Technology


I'm writing a benchmark to compare different technologies and their performance accross various platforms (on Linux). One of the platform is an Intel Broadwell-H  (Core i7-5775C with integrated GPU Iris Pro 6200), so I'm testing the various ways to offload a code on my GPU using Cilk Plus. Right now, I'm trying to use SVM so I followed this tutorial but I'm facing some problems. Here's my code :

Cilk Fibonancci doesn't work properly

I get the Fibonancci example from a web site but the output doesn't make any sense as serial code takes 2.25 sec while the parallel code takes 4.5 sec.
I'm using visual studio 2013, and Intel parallel studio
this is the code:

#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#include <cilk/cilk.h>
#include <cilk/cilk_api.h>

Subscribe to Intel® Cilk™ Plus