Home
last modified time | relevance | path

Searched refs:replicated (Results 1 – 25 of 121) sorted by relevance

12345

/external/tensorflow/tensorflow/core/api_def/base_api/
Dapi_def_TPUReplicatedOutput.pbtxt4 summary: "Connects N outputs from an N-way replicated TPU computation."
6 This operation holds a replicated output from a `tpu.replicate()` computation subgraph.
7 Each replicated output has the same shape and type alongside the input.
14 The above computation has a replicated output of two replicas.
Dapi_def_TPUReplicatedInput.pbtxt4 summary: "Connects N inputs to an N-way replicated TPU computation."
6 This operation holds a replicated input to a `tpu.replicate()` computation subgraph.
7 Each replicated input has the same shape and type alongside the output.
16 The above computation has a replicated input of two replicas.
Dapi_def_TPUPartitionedInput.pbtxt19 those inputs are replicated.
Dapi_def_CollectivePermute.pbtxt29 summary: "An Op to permute tensors across replicated TPU instances."
Dapi_def_TPUReplicateMetadata.pbtxt46 summary: "Metadata indicating how the TPU computation should be replicated."
Dapi_def_CrossReplicaSum.pbtxt30 summary: "An Op to sum inputs across replicated TPU instances."
Dapi_def_Tile.pbtxt19 and the values of `input` are replicated `multiples[i]` times along the 'i'th
/external/tensorflow/tensorflow/compiler/mlir/tensorflow/tests/
Dtf_device_ops_invalid.mlir13 // Check that a replicate replicated inputs where operand sizes do not match
17 // expected-error@-1 {{'tf_device.replicate' expects number of operands for replicated input 0 to b…
129 // Check number of replicated inputs is evenly divisible by 'n'.
132 // expected-error@-1 {{'tf_device.replicate' op expects number of replicated inputs (4) to be evenl…
140 // Check number of replicated inputs / 'n' + number of packed inputs matches the
144 …ate' op expects number of block arguments (2) to be equal to number of replicated inputs (3) / 'n'…
152 // Check that a replicate with incompatible replicated operand and block
Dtpu-dynamic-layout-pass.mlir3 // Tests that the pass can transform non-replicated execution.
235 // Tests that the pass can transform replicated execution.
237 // CHECK: func @replicated(%[[ARG0:.*]]: tensor<*x!tf.resource> {tf.device = "/device:CPU:0"}) -> t…
238 func @replicated(%arg0: tensor<*x!tf.resource> {tf.device = "/device:CPU:0"}) -> tensor<i32> {
280 // Tests that the pass can transform replicated execution with packed inputs.
317 // Tests that the pass can transform replicated execution with both replicated
320 // CHECK: func @replicated(%[[ARG0:.*]]: tensor<*x!tf.resource> {tf.device = "/device:CPU:0"}) -> t…
321 func @replicated(%arg0: tensor<*x!tf.resource> {tf.device = "/device:CPU:0"}, %arg1: tensor<*x!tf.r…
Dannotate-parameter-replication.mlir3 // Tests that an operand from outside the replicated region is annotated.
68 // Tests that a non-replicated ClusterFuncOp is not annotated.
/external/tensorflow/tensorflow/compiler/mlir/tensorflow/tests/compile_mlir_util/
Dargument-sharding.mlir37 // CHECK-SAME: {replicated}
44 // CHECK-SAME: sharding={replicated}
Dresult-sharding.mlir38 // CHECK-SAME: {replicated}
/external/mesa3d/src/panfrost/util/
Dpan_lower_framebuffer.c160 nir_ssa_def *replicated[4]; in pan_pack_pure_32() local
163 replicated[i] = nir_channel(b, v, i % v->num_components); in pan_pack_pure_32()
165 return nir_vec(b, replicated, 4); in pan_pack_pure_32()
180 nir_ssa_def *replicated[4]; in pan_pack_pure_16() local
190 replicated[i] = nir_pack_32_2x16(b, nir_vec(b, parts, 2)); in pan_pack_pure_16()
193 return nir_vec(b, replicated, 4); in pan_pack_pure_16()
252 nir_ssa_def *replicated[4] = { v, v, v, v }; in pan_replicate_4() local
253 return nir_vec(b, replicated, 4); in pan_replicate_4()
/external/tensorflow/tensorflow/compiler/xla/service/
Dhlo_sharding.h292 explicit HloSharding(bool manual, bool replicated, in HloSharding() argument
294 : replicated_(replicated), in HloSharding()
295 maximal_(replicated), in HloSharding()
Dhlo_sharding_test.cc104 auto* replicated = proto.add_tuple_shardings(); in TEST_F() local
105 replicated->set_type(OpSharding::REPLICATED); in TEST_F()
106 *replicated->add_metadata() = GetMetadata("c"); in TEST_F()
/external/tensorflow/tensorflow/compiler/mlir/tensorflow/ir/
Dtf_device_ops.td180 let summary = "Wraps an N-way replicated computation.";
183 The region held by this operation represents a computation that is replicated
187 from the outer scope. The device name map specifies devices on which replicated
197 the associated replicated device from `devices` if the tf_device.launch refers
201 Operands are replicated inputs and packed inputs.
205 the operands are matching in order the `devices` attribute. Each replicated
210 Operands not replicated can be implicitly captured by ops in the region. Results
211 are replicated each from the regions terminator.
/external/llvm-project/llvm/test/CodeGen/SystemZ/
Dvec-move-19.ll3 ; Test that a loaded value which is replicated is not inserted also in any
Dvec-const-13.ll139 ; ...and again with the lower bits of the replicated constant.
161 ; ...and again with the lower bits of the replicated constant.
/external/llvm-project/mlir/test/Conversion/StandardToLLVM/
Dconvert-argattrs.mlir4 // When expanding the memref to multiple arguments, argument attributes are replicated.
/external/tensorflow/tensorflow/core/protobuf/tpu/
Dcompile_metadata.proto78 // replica here. Reconsider when replicated model-parallelism is implemented
119 // Enables use of XLA collectives for broadcast of replicated parameters to
/external/libavc/common/arm/
Dih264_padding_neon.s51 @* The top row of a 2d array is replicated for pad_size times at the top
120 @* The left column of a 2d array is replicated for pad_size times at the left
256 @* The left column of a 2d array is replicated for pad_size times at the left
384 @* The right column of a 2d array is replicated for pad_size times at the right
530 @* The right column of a 2d array is replicated for pad_size times at the right
/external/tensorflow/tensorflow/compiler/mlir/tensorflow/transforms/
Dtpu_rewrite_pass.cc506 const bool replicated = tpu_devices.size() != 1; in BuildParallelExecuteOp() local
529 std::string device = replicated in BuildParallelExecuteOp()
547 const bool replicated = tpu_devices.size() != 1; in AssignDevicesToReplicatedExecute() local
550 std::string device = replicated ? tensorflow::GetDeviceAliasForLogicalCore(0) in AssignDevicesToReplicatedExecute()
Dtf_passes.td139 inputs and outputs to and from a replicated TPU computation. The number of times
140 a TPU computation is replicated is defined in the `tf.TPUReplicateMetadata` op
150 computation is replicated (`num_replicas` > 1), the `num_replicas` attribute is
158 For example, the following non replicated computation:
184 The following replicated computation:
506 let summary = "Reorder replicated and partitioned input ops.";
609 For example, a non replicated `tf_device.cluster_func`:
638 A replicated `tf_device.cluster_func`:
672 A non replicated `tf_device.cluster_func` with the model parallelism:
/external/llvm/test/CodeGen/SystemZ/
Dvec-const-13.ll139 ; ...and again with the lower bits of the replicated constant.
161 ; ...and again with the lower bits of the replicated constant.
/external/javaparser/javaparser-core-testing-bdd/src/test/resources/com/github/javaparser/
Dvisitor_scenarios.story1 Scenario: A class that is replicated using a CloneVisitor should be equal to the source

12345