d30c0221cf
* Split ops into X_graph variants as discussed; * Remove tokens from non-Graph region variants and rely on side-effect modelling there while removing side-effect modelling from Graph variants and relying on explicit ordering there; * Make tokens required to be produced by Graph variants - but kept explicit token type specification given previous discussion on this potentially being configurable in future; This results in duplicating some code. I considered adding helper functions but decided against adding an abstraction there early given size of duplication and creating accidental coupling. Differential Revision: https://reviews.llvm.org/D127813
57 lines
2.4 KiB
MLIR
57 lines
2.4 KiB
MLIR
// RUN: mlir-opt %s --allow-unregistered-dialect | mlir-opt --allow-unregistered-dialect | FileCheck %s
|
|
// RUN: mlir-opt %s --allow-unregistered-dialect --mlir-print-op-generic | mlir-opt --allow-unregistered-dialect | FileCheck %s
|
|
|
|
// CHECK-LABEL: ml_program.func private @extern_func
|
|
ml_program.func private @extern_func(i32) -> i32
|
|
|
|
// CHECK-LABEL: ml_program.func @defined_func
|
|
ml_program.func @defined_func(%arg0 : i32) -> i32 {
|
|
ml_program.return %arg0 : i32
|
|
}
|
|
|
|
// CHECK-LABEL: ml_program.subgraph private @extern_subgraph
|
|
ml_program.subgraph private @extern_subgraph(i32) -> i32
|
|
|
|
// CHECK-LABEL: ml_program.subgraph @compute_subgraph
|
|
ml_program.subgraph @compute_subgraph(%arg0 : i32) -> i32 {
|
|
%token = ml_program.token
|
|
%1 = "unregistered.dummy"(%0, %token) : (i32, !ml_program.token) -> i32
|
|
%0 = "unregistered.dummy"(%arg0) : (i32) -> i32
|
|
ml_program.output %0 : i32
|
|
}
|
|
|
|
// CHECK: ml_program.global private @global_same_type(dense<4> : tensor<4xi32>) : tensor<4xi32>
|
|
ml_program.global private @global_same_type(dense<4> : tensor<4xi32>) : tensor<4xi32>
|
|
|
|
// CHECK: ml_program.global private mutable @global_mutable_undef : tensor<?xi32>
|
|
ml_program.global private mutable @global_mutable_undef : tensor<?xi32>
|
|
|
|
// CHECK: ml_program.global private mutable @global_extern(#extern) : tensor<?xi32>
|
|
ml_program.global private mutable @global_extern(#ml_program.extern : tensor<4xi32>) : tensor<?xi32>
|
|
|
|
// CHECK-LABEL: @global_load_const
|
|
ml_program.func @global_load_const() -> tensor<4xi32> {
|
|
%0 = ml_program.global_load_const @global_same_type : tensor<4xi32>
|
|
ml_program.return %0 : tensor<4xi32>
|
|
}
|
|
|
|
// CHECK-LABEL: @global_load_store
|
|
ml_program.func @global_load_store() {
|
|
%0 = ml_program.global_load @global_mutable_undef : tensor<?xi32>
|
|
ml_program.global_store @global_mutable_undef = %0 : tensor<?xi32>
|
|
ml_program.return
|
|
}
|
|
|
|
// CHECK-LABEL: @global_load_store_tokens
|
|
ml_program.subgraph @global_load_store_tokens() -> (tensor<?xi32>, !ml_program.token) {
|
|
%token1 = ml_program.token
|
|
%0, %token2 = ml_program.global_load_graph @global_mutable_undef
|
|
ordering(() -> !ml_program.token) : tensor<?xi32>
|
|
%token3 = ml_program.global_store_graph @global_mutable_undef = %0
|
|
ordering(%token1, %token2 -> !ml_program.token) : tensor<?xi32>
|
|
%token4 = ml_program.global_store_graph @global_mutable_undef = %0
|
|
ordering(%token3 -> !ml_program.token) : tensor<?xi32>
|
|
|
|
ml_program.output %0, %token3 : tensor<?xi32>, !ml_program.token
|
|
}
|