I’m trying to run this tutorial quick-start-tutorial-for-compiling-deep-learning-models and I can run all the steps successfully with target = "llvm".
However I met the following error when I changed the target to tvm.target.intel_graphics():
Check failed: analyzer.CanProve(call->args[3] == call->args[4]) == false: Intel warp shuffle dose not support width != warp_size
Anyone know why?