Intel® MPI Library
Get help with building, analyzing, optimizing, and scaling high-performance computing (HPC) applications.
2158 Discussions

What hardware will Intel Ct technology support

Laszlo_Kornyei
Beginner
420 Views

Dear Community,

My research group in Hungary is investigating new tools for high performance computing. So I came across Rapidmind, and its aquisition by Intel. As far as I know, Rapidmind aimed at transparent development tools for multiple HPC hardware, like clusters, GPU, etc. I am curious, which type of hardware will Intel Ct technology support? I am aware of support for multi-threaded multi-core Intel CPUs. Will there be support for other type systems, like (eg. nvidia or ati) GPU-s? Or for cluster machines? Or even special hardware like FPGA? I wonder, if I develop code with Intel Ct, will I be able to compile it to these platforms?

I am sorry, if this is not the adequate place to ask this question, but I did not find any clear information on this.

Thank You,

Laszlo Kornyei

0 Kudos
2 Replies
TimP
Honored Contributor III
420 Views
It seems unlikely that anyone has full answers for your questions at this stage. The public Ct beta test for VS2008 only (plus TBB, Intel C++ preferred) has been announced for next month. That apparently doesn't yet involve full integration of the Intel and RapidMind versions. According to the announcements, the plans involve going ahead with RapidMind projects and full reconciliation of Intel and RapidMind.
At a slightly higher level (at least to the eye of those who use OpenMP), influential members of the OpenMP community have announced intention to expand the standard so as to cover GPGPUs and extend the use of hybrid OpenMP/MPI. So it will be interesting to see how the tools develop.
0 Kudos
Gergana_S_Intel
Employee
420 Views
Quoting - Laszlo Kornyei

I am sorry, if this is not the adequate place to ask this question, but I did not find any clear information on this.

Hi Laszlo,

If you indeed have any questions regarding to Intel's Ct Technology, I would point you to visit their data parallelism website.

Regards,
~Gergana

0 Kudos
Reply