https://community.intel.com/t5/OpenCL-for-CPU/Too-imprecise-native-sin-native-cos-implementation/m-p/972941#M2409 <P>Running my OpenCL app on Intel GPUs I encountered precision issue. Results were too imprecise and often failed to pass validation vs CPU-based (or AMD/nVidia GPU based) ones.</P> <P>Considerable improvement was only when I refused from native_cos/native_sin in oclFFT &nbsp;(Apple's FFT release for OpenCL) and in other parts of app code.</P> <P>I have no such issues with same OpenCL code on GPUs from other vendors (both AMD and nVidia).</P> <P>Please consider some improvements for native trigonometric function presion in future SDK/hardware releases. Too low precision makes them just un-useful no matter how speedy they are. Make precision at least in line with other vendors.</P> <P></P> Sat, 02 Nov 2013 09:46:37 GMT Raistmer 2013-11-02T09:46:37Z https://community.intel.com/t5/OpenCL-for-CPU/Too-imprecise-native-sin-native-cos-implementation/m-p/972941#M2409 <P>Running my OpenCL app on Intel GPUs I encountered precision issue. Results were too imprecise and often failed to pass validation vs CPU-based (or AMD/nVidia GPU based) ones.</P> <P>Considerable improvement was only when I refused from native_cos/native_sin in oclFFT &nbsp;(Apple's FFT release for OpenCL) and in other parts of app code.</P> <P>I have no such issues with same OpenCL code on GPUs from other vendors (both AMD and nVidia).</P> <P>Please consider some improvements for native trigonometric function presion in future SDK/hardware releases. Too low precision makes them just un-useful no matter how speedy they are. Make precision at least in line with other vendors.</P> <P></P> Sat, 02 Nov 2013 09:46:37 GMT https://community.intel.com/t5/OpenCL-for-CPU/Too-imprecise-native-sin-native-cos-implementation/m-p/972941#M2409 Raistmer 2013-11-02T09:46:37Z https://community.intel.com/t5/OpenCL-for-CPU/Too-imprecise-native-sin-native-cos-implementation/m-p/972942#M2410 <P>Native functions should be faster but not necessarily precise. I am not sure how much difference is acceptable. Are you seeing precision issues with native_sin and native_cos only? Thanks for the feedback.</P> <P>Raghu</P> Tue, 05 Nov 2013 00:46:24 GMT https://community.intel.com/t5/OpenCL-for-CPU/Too-imprecise-native-sin-native-cos-implementation/m-p/972942#M2410 Raghupathi_M_Intel 2013-11-05T00:46:24Z