ghostplant opened a new issue #4547: CUDA source emited from hybrid script contains undefined `vthread` URL: https://github.com/apache/incubator-tvm/issues/4547 ```cpp extern "C" __global__ void test_kernel0( float* __restrict__ outer_product, float* __restrict__ A) { for (int i_inner_inner_inner = 0; i_inner_inner_inner < 256; ++i_inner_inner_inner) { outer_product[((((((int)blockIdx.x) * 65536) + (((int)vthread) * 32768)) + (((int)threadIdx.x) * 256)) + i_inner_inner_inner)] = A[((((((int)blockIdx.x) * 65536) + (((int)vthread) * 32768)) + (((int)threadIdx.x) * 256)) + i_inner_inner_inner)]; } } ``` Its lower code is: ``` produce test_op { // attr [0] extern_scope = 0 // attr [iter_var(blockIdx.x, , blockIdx.x)] thread_extent = 16384 // attr [iter_var(vthread, , vthread)] thread_extent = 2 // attr [iter_var(threadIdx.x, , threadIdx.x)] thread_extent = 128 for (i.inner.inner.inner, 0, 256) { outer_product[((((blockIdx.x*65536) + (vthread*32768)) + (threadIdx.x*256)) + i.inner.inner.inner)] = A[((((blockIdx.x*65536) + (vthread*32768)) + (threadIdx.x*256)) + i.inner.inner.inner)] } } ``` TVM backend version: v0.6
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
