TR Forums

With all of the new products coming out from AMD and Intel will we be seeing new compliers for each system? Will they try and take advatange of the new capabilities of each chipset? I.E. GPUs are inheriently suited for multi-threaded tasks. Some OS tasks might run faster on the APU then on the CPU in the new chips. Where will we see the true benefits of these new Architectures? Do you see it being more in the instruction set on each architecture, the compliers available for each, or the libraries on each? Will they be shared or will they start to be independent of each other since the APUs are being integrated onto the CPUs? Does anyone see another OpenGL vs DirectX battle oncoming?

I'm starting this topic just to talk in general about the new architectures coming from AMD and Intel and how they will relate to the software that runs on them.

Intel's had their ICC compiler for a while now which whilst the code will run on any x86, it checks for an Intel processor and only runs the best code on Genuine Intel processors. If you hack the binary then the AMD processors can get a boost. Apart from that, I see GCC in use in a lot of places and they keep tweaking the code generated for using the newer processor architectures. I believe most processor vendors contribute to GCC to help the code generation.