CINN
CINN copied to clipboard
Compiler Infrastructure for Neural Networks
support computeAt when forloops' range is consistent
https://github.com/PaddlePaddle/CINN/pull/429
https://github.com/PaddlePaddle/CINN/pull/417 computeAt lowered ir is different with computeAt2 lowered ir
添加Instrucionts及相关测试 运行`test_cinn_instruction`测试即可看到demo结果
schedule分别设置computeAt和computeAt2, default case生成的ir不同 @haozech
isl for1 wrong elimination after splitting 执行test02_matmul_case i_outer和j_outer是for1,isl错误消除,导致最后一个轴vectorize根据level找不到for轴
Build a framework for developers to conveniently compare CINN's op performance with TVM's op performance.