junrushao1994 commented on a change in pull request #10043: URL: https://github.com/apache/tvm/pull/10043#discussion_r791080656
########## File path: python/tvm/meta_schedule/schedule_rule/multi_level_tiling.py ########## @@ -0,0 +1,88 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. +"""Multi-level tiling with reuse.""" +from typing import Any, Dict, List, NamedTuple, Optional + +from tvm._ffi import register_object + +from .. import _ffi_api +from .schedule_rule import ScheduleRule + + +class ReuseType(NamedTuple): + """Reuse type.""" + + req: str + levels: List[int] + scope: str + + def as_dict(self) -> Dict[str, Any]: + """Return the dict representation of the reuse type.""" + return { + "req": self.req, + "levels": self.levels, + "scope": self.scope, + } + + +@register_object("meta_schedule.MultiLevelTiling") +class MultiLevelTiling(ScheduleRule): + """Multi-level tiling with reuse. + + Parameters + ---------- + structure : str + The tiling structure. Recommended: + - 'SSRSRS' on CPU + - 'SSSRRSRS' on GPU + tile_bind : Optional[List[str]] + For each level of tiles, which thread axis it is bound to. Recommended: + - None on CPU + - [blockIdx.x, vthread.x, threadIdx.x] on GPU + use_tensor_core : bool + Whether to apply tensor core wmma intrinsic for the computation Review comment: Yes. We want to support auto tensorization for all hardware platforms. use-tensor-core is probably not a good name here - maybe use-tensor-intrin could be a better one ########## File path: python/tvm/meta_schedule/schedule_rule/multi_level_tiling.py ########## @@ -0,0 +1,88 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. +"""Multi-level tiling with reuse.""" +from typing import Any, Dict, List, NamedTuple, Optional + +from tvm._ffi import register_object + +from .. import _ffi_api +from .schedule_rule import ScheduleRule + + +class ReuseType(NamedTuple): + """Reuse type.""" + + req: str + levels: List[int] + scope: str + + def as_dict(self) -> Dict[str, Any]: + """Return the dict representation of the reuse type.""" + return { + "req": self.req, + "levels": self.levels, + "scope": self.scope, + } + + +@register_object("meta_schedule.MultiLevelTiling") +class MultiLevelTiling(ScheduleRule): + """Multi-level tiling with reuse. + + Parameters + ---------- + structure : str + The tiling structure. Recommended: + - 'SSRSRS' on CPU + - 'SSSRRSRS' on GPU + tile_bind : Optional[List[str]] + For each level of tiles, which thread axis it is bound to. Recommended: + - None on CPU + - [blockIdx.x, vthread.x, threadIdx.x] on GPU + use_tensor_core : bool + Whether to apply tensor core wmma intrinsic for the computation Review comment: Well, you are right. We might want to remove all the tensor core hard-coding here and focusing on AutoScheduler alignment first during upstreaming -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
