mbs-octoml commented on code in PR #11474: URL: https://github.com/apache/tvm/pull/11474#discussion_r887412903
########## tests/python/relay/transform/test_compiler_function_utils.py: ########## @@ -0,0 +1,162 @@ +# 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 +"""Unit tests for the OutlineCompilerFunctionsWithExistingGlobalSymbols and + MarkCompilerFunctionsAsExtern external codegen helper passes.""" + +import tvm +import tvm.testing +import numpy as np + + +def make_const(dtype, shape): + return tvm.relay.const(np.random.rand(*shape).astype(dtype)) + + +def make_consts(dtype, shapes): + return [make_const(dtype, shape) for shape in shapes] + + +metatable = { + "relay.Constant": make_consts( + "float16", + [ + (2304, 768), # 0 + (2304,), # 1 + (600, 32, 64), # 2 + ], + ), + "attributes": [{"relay_attrs": None}], +} + + +def inlined_mod(): + return tvm.parser.parse( + """ + #[version = "0.0.5"] + def @main(%x0 : Tensor[(1600, 768), float16], %x3 : Tensor[(600, 32, 64), float16]) -> (Tensor[(1600, 2304), float16], Tensor[(600, 32, 32), float16]) { + %0 = fn(%y_0_i0: Tensor[(1600, 768), float16], %y_0_i1: Tensor[(2304, 768), float16], %y_0_i2: Tensor[(2304), float16], + Inline=1, Compiler="cutlass", global_symbol="tvmgen_default_cutlass_main_0", Primitive=1) -> Tensor[(1600, 2304), float16] { + %4 = fn (%FunctionVar_0_0: Tensor[(1600, 768), float16], %FunctionVar_0_1: Tensor[(2304, 768), float16], %FunctionVar_0_2: Tensor[(2304), float16], + PartitionedFromPattern="nn.dense_add_", Composite="cutlass.dense_bias") -> Tensor[(1600, 2304), float16] { + %5 = nn.dense(%FunctionVar_0_0, %FunctionVar_0_1, units=2304); + add(%5, %FunctionVar_0_2) + }; + %4(%y_0_i0, %y_0_i1, %y_0_i2) + }; + %1 = %0(%x0, meta[relay.Constant][0], meta[relay.Constant][1]); + %2 = fn(%y_3_i0: Tensor[(600, 32, 64), float16], %y_3_i1: Tensor[(600, 32, 64), float16], + Inline=1, Compiler="cublas", global_symbol="tvmgen_default_cublas_main_3", Primitive=1) -> Tensor[(600, 32, 32), float16] { + %6 = fn (%FunctionVar_0_01: Tensor[(600, 32, 64), float16], %FunctionVar_0_11: Tensor[(600, 32, 64), float16], + PartitionedFromPattern="nn.batch_matmul_", Composite="cublas.batch_matmul") -> Tensor[(600, 32, 32), float16] { + nn.batch_matmul(%FunctionVar_0_01, %FunctionVar_0_11, out_dtype="float16", transpose_b=True) + }; + %6(%y_3_i0, %y_3_i1) + }; + %3 = %2(%x3, meta[relay.Constant][2]); + (%1, %3) + } + """, + "from_string", + None, + metatable, + ) + + +def expected_outlined_mod(): + return tvm.parser.parse( + """ + #[version = "0.0.5"] + def @main(%x0 : Tensor[(1600, 768), float16], %x3 : Tensor[(600, 32, 64), float16]) -> (Tensor[(1600, 2304), float16], Tensor[(600, 32, 32), float16]) { + %1 = @tvmgen_default_cutlass_main_0(%x0, meta[relay.Constant][0], meta[relay.Constant][1]); + %2 = fn(%y_3_i0: Tensor[(600, 32, 64), float16], %y_3_i1: Tensor[(600, 32, 64), float16], + Inline=1, Compiler="cublas", global_symbol="tvmgen_default_cublas_main_3", Primitive=1) -> Tensor[(600, 32, 32), float16] { + %6 = fn (%FunctionVar_0_01: Tensor[(600, 32, 64), float16], %FunctionVar_0_11: Tensor[(600, 32, 64), float16], + PartitionedFromPattern="nn.batch_matmul_", Composite="cublas.batch_matmul") -> Tensor[(600, 32, 32), float16] { + nn.batch_matmul(%FunctionVar_0_01, %FunctionVar_0_11, out_dtype="float16", transpose_b=True) + }; + %6(%y_3_i0, %y_3_i1) + }; + %3 = %2(%x3, meta[relay.Constant][2]); + (%1, %3) + } + + def @tvmgen_default_cutlass_main_0(%y_0_i0: Tensor[(1600, 768), float16], %y_0_i1: Tensor[(2304, 768), float16], %y_0_i2: Tensor[(2304), float16], + Inline=1, Compiler="cutlass", global_symbol="tvmgen_default_cutlass_main_0", Primitive=1) -> Tensor[(1600, 2304), float16] { + %4 = fn (%FunctionVar_0_0: Tensor[(1600, 768), float16], %FunctionVar_0_1: Tensor[(2304, 768), float16], %FunctionVar_0_2: Tensor[(2304), float16], + PartitionedFromPattern="nn.dense_add_", Composite="cutlass.dense_bias") -> Tensor[(1600, 2304), float16] { + %5 = nn.dense(%FunctionVar_0_0, %FunctionVar_0_1, units=2304); + add(%5, %FunctionVar_0_2) + }; + %4(%y_0_i0, %y_0_i1, %y_0_i2) + } + """, + "from_string", + None, + metatable, + ) + + +def expected_extern_mod(): + return tvm.parser.parse( + """ + #[version = "0.0.5"] + def @main(%x0 : Tensor[(1600, 768), float16], %x3 : Tensor[(600, 32, 64), float16]) -> (Tensor[(1600, 2304), float16], Tensor[(600, 32, 32), float16]) { + %1 = call_lowered(@tvmgen_default_cutlass_main_0, (%x0, meta[relay.Constant][0], meta[relay.Constant][1]), metadata=meta[attributes][0]); + %2 = fn(%y_3_i0: Tensor[(600, 32, 64), float16], %y_3_i1: Tensor[(600, 32, 64), float16], + Inline=1, Compiler="cublas", global_symbol="tvmgen_default_cublas_main_3", Primitive=1) -> Tensor[(600, 32, 32), float16] { + %6 = fn (%FunctionVar_0_01: Tensor[(600, 32, 64), float16], %FunctionVar_0_11: Tensor[(600, 32, 64), float16], + PartitionedFromPattern="nn.batch_matmul_", Composite="cublas.batch_matmul") -> Tensor[(600, 32, 32), float16] { + nn.batch_matmul(%FunctionVar_0_01, %FunctionVar_0_11, out_dtype="float16", transpose_b=True) + }; + %6(%y_3_i0, %y_3_i1) + }; + %3 = %2(%x3, meta[relay.Constant][2]); + (%1, %3) + } + + def @tvmgen_default_cutlass_main_0(%y_0_i0: Tensor[(1600, 768), float16], %y_0_i1: Tensor[(2304, 768), float16], %y_0_i2: Tensor[(2304), float16], + Extern=1) -> Tensor[(1600, 2304), float16] { + %4 = fn (%FunctionVar_0_0: Tensor[(1600, 768), float16], %FunctionVar_0_1: Tensor[(2304, 768), float16], %FunctionVar_0_2: Tensor[(2304), float16], + PartitionedFromPattern="nn.dense_add_", Composite="cutlass.dense_bias") -> Tensor[(1600, 2304), float16] { Review Comment: The "Primitive=1" attribute is a generic way to indicate 'this function represents a sub-graph which is to be lowered/compiled as a unit'. The additional "Compiler=foo" attributes indicates foo is the one doing the compiling, otherwise it's TVM. That kinda makes sense. The "PartitionedFromPattern" and "Composite" duplication seems unnecessary to me. Why not just use "PartitionedFromPattern"? The latter is always exactly equal to the former as per merge_composite.cc. However, note there's nesting here: - the partition itself is represented using Primitive=1, Compiler=foo. That's groups what we want to hand off. - the body of that function may have calls to literal functions with "PartitionedFromPattern=foo.bar". That groups sub-expressions into translation units. That threw me off for a while :-) -- 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]
