By: Rohit (.delete@this..), August 15, 2011 5:52 pm
Room: Moderated Discussions
Mark Roulo (nothanks@xxx.com) on 8/15/11 wrote:
---------------------------
>My expectation is that OpenCL performance on x86 is a *long* way from being performance
>competitive with hand coded SSE/AVX intrinsics. Intel's marketing question would
>be, "Is there a large enough market for performance willing to code in OpenCL, but unwilling to use intrinsics?"
Intel did a presentation a while ago where they claimed OCL gave 95% of hand coded intrinsics code for n body.
Obviously, it won't translate to all problems, but the difference shouldn't be substantial, especially, if you compare the productivity of OCL vs hand coded intrinsics.
---------------------------
>My expectation is that OpenCL performance on x86 is a *long* way from being performance
>competitive with hand coded SSE/AVX intrinsics. Intel's marketing question would
>be, "Is there a large enough market for performance willing to code in OpenCL, but unwilling to use intrinsics?"
Intel did a presentation a while ago where they claimed OCL gave 95% of hand coded intrinsics code for n body.
Obviously, it won't translate to all problems, but the difference shouldn't be substantial, especially, if you compare the productivity of OCL vs hand coded intrinsics.