Article: Parallelism at HotPar 2010
By: hobold (hobold.delete@this.vectorizer.org), August 3, 2010 4:41 am
Room: Moderated Discussions
Gabriele Svelto (gabriele.svelto@gmail.com) on 8/3/10 wrote:
---------------------------
[...]
>The reason is that with an IGP you will basically have a wide vector processor
>hooked to your main memory pool which guarantees quick communication and little
>overhead (with the appropriate abstractions you might even be able to reuse the
>same memory buffers without having to copy them around).
That's why the immediate future of computational devices is so interesting. Intel tries to evolve CPUs more towards GPUs. Nvidia tries to push GPUs towards CPUs. And AMD might not have the most sophisticated CPU or GPU, but by far the best chance to integrate the two really well.
I find it impossible to predict which of the three approaches is going to win. I don't see clear advantages or disadvantages to either of them.
---------------------------
[...]
>The reason is that with an IGP you will basically have a wide vector processor
>hooked to your main memory pool which guarantees quick communication and little
>overhead (with the appropriate abstractions you might even be able to reuse the
>same memory buffers without having to copy them around).
That's why the immediate future of computational devices is so interesting. Intel tries to evolve CPUs more towards GPUs. Nvidia tries to push GPUs towards CPUs. And AMD might not have the most sophisticated CPU or GPU, but by far the best chance to integrate the two really well.
I find it impossible to predict which of the three approaches is going to win. I don't see clear advantages or disadvantages to either of them.