On 2018-03-19 01:49 AM, Sergey Sharybin wrote: > Hi, > > You should be able to fix this compilation error by changing it to: return > _mm_castsi128_ps(_mm_shuffle_epi32(a, _MM_SHUFFLE(i3, i2, i1, i0))); > > Patch version is: https://hastebin.com/efopofoboy.diff > > Let me know if it indeed fixes the problem. > > On Sat, Mar 17, 2018 at 8:02 PM, Luya Tshimbalanga <[email protected]> > wrote: > >> On 2018-03-17 12:32 AM, PerfectionCat wrote: >>> Hi there. >>> >>> /builddir/build/BUILD/blender-2.79a/intern/cycles/kernel/../util/util_sseb.h:119:57: >> error: could not convert '_mm_shuffle_epi32(a, ((((i3 << 6) | (i2 << 4)) | >> (i1 << 2)) | i0))' from '__m128i' {aka '__vector(2) long long int'} to >> 'const ccl::sseb' return _mm_shuffle_epi32(a, _MM_SHUFFLE(i3, i2, i1, i0)); >> ^ >>> make[2]: *** >>> [intern/cycles/kernel/CMakeFiles/cycles_kernel.dir/build.make:90: >> intern/cycles/kernel/CMakeFiles/cycles_kernel.dir/kernels/cpu/kernel_sse2.cpp.o] >> Error 1 >>> >> I don't know how to resolved that issue as debugging such codes is >> beyond my skills. It seems the rules arestricter in GCC 8 than GCC7. >> Could someone provide a patch for testing purpose? >> >> Thanks >> >> >> -- >> Luya Tshimbalanga >> Graphic & Web Designer >> E: [email protected] >> W: http://www.coolest-storm.net >> >> _______________________________________________ >> Bf-committers mailing list >> [email protected] >> https://lists.blender.org/mailman/listinfo/bf-committers >> The patch resolved one issue but another came out as seen on build.log below: https://kojipkgs.fedoraproject.org//work/tasks/8115/25818115/build.log
Full build system on: https://koji.fedoraproject.org/koji/taskinfo?taskID=25818115 And available architectures: https://koji.fedoraproject.org/koji/taskinfo?taskID=25818095 -- Luya Tshimbalanga Graphic & Web Designer E: [email protected] W: http://www.coolest-storm.net _______________________________________________ Bf-committers mailing list [email protected] https://lists.blender.org/mailman/listinfo/bf-committers
