https://community.intel.com/t5/Intel-Integrated-Performance/ippiRemap-issue/m-p/876146#M9557 No I don't get a pure copy, with the 5.3.1 release.<BR /><BR />I'm using 32x32 images with cubic interpolation using the dll's on a coro duo processor.<BR /> Sun, 23 Dec 2007 20:16:17 GMT darin1 2007-12-23T20:16:17Z https://community.intel.com/t5/Intel-Integrated-Performance/ippiRemap-issue/m-p/876145#M9556 With the latest version if ipp 5.3.1.064<BR /><BR />I find that I have different results now with ippiRemap. For example passing in the identity pxMap and pyMap so that the following should be occurring dst[x,y] = src[x,y] I'm finding that the border of the dst are not the correct values, for example dst[0,0] != src[0,0] while for values in the interior appear to be correct.<BR /><BR />While with the previous version of ipp 5.2 this function seems to work as expected, where using the identity maps for pxMap, pyMap the src == dst for all pels.<BR /><BR /><BR /> Tue, 18 Dec 2007 21:21:32 GMT https://community.intel.com/t5/Intel-Integrated-Performance/ippiRemap-issue/m-p/876145#M9556 darin1 2007-12-18T21:21:32Z https://community.intel.com/t5/Intel-Integrated-Performance/ippiRemap-issue/m-p/876146#M9557 No I don't get a pure copy, with the 5.3.1 release.<BR /><BR />I'm using 32x32 images with cubic interpolation using the dll's on a coro duo processor.<BR /> Sun, 23 Dec 2007 20:16:17 GMT https://community.intel.com/t5/Intel-Integrated-Performance/ippiRemap-issue/m-p/876146#M9557 darin1 2007-12-23T20:16:17Z