I'm running a simple test case that makes me question my understanding of how varying types are created when varying pointers to varying types are dereferenced. My understanding was that each pointer would be dereferenced with the appropriate lane. But this appears to not be the case.
Consider: export void run() { varying int p[programCount]; for (uniform int i = 0; i < programCount; ++i) { p[i] = i * programCount + programIndex; print("%\n", p[i]); } varying int * uniform a = &p[0]; print("%\n", a); varying int * varying av = a; print("%\n", av); varying int aDeref = *a; varying int avDeref = *av; print("%\n", aDeref); print("%\n", avDeref); } Compiled with ispc-1.9.2 as: ispc test.ispc --target=sse2-i32x4 -o test_ispc.o -h test_ispc.h The following output is produced: [0,1,2,3] [4,5,6,7] [8,9,10,11] [12,13,14,15] 0x7ffd8568d240 [0x7ffd8568d240,0x7ffd8568d240,0x7ffd8568d240,0x7ffd8568d240] [0,1,2,3] [0,0,0,0] The avDeref value suprises me. I was expecting [0,1,2,3]. In other words, the pointer in lane 0 would be dereferenced for lane 0, the pointer in lane 1 would be dereferenced at lane 1, etc... But instead all of them seem to be dereferenced for lane 0. I've tested this with avx1-i32x8 and avx2-i32x8 and seen similar results. So, what am I misunderstanding about varying pointers to varying types? Thanks, -Brian -- You received this message because you are subscribed to the Google Groups "Intel SPMD Program Compiler Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to ispc-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.