Home
last modified time | relevance | path

Searched refs:hlo_module_config_ (Results 1 – 10 of 10) sorted by relevance

/external/tensorflow/tensorflow/compiler/xla/service/cpu/
Ddot_op_emitter.cc177 return options::LlvmIrGemvTilingFactor(hlo_module_config_) in GetGemvTilingFactor()
188 return options::LlvmIrGemmTileSize(hlo_module_config_) in GetGemmTileSize()
200 const HloModuleConfig& hlo_module_config_; member in xla::cpu::__anond13f39450111::DotOpEmitter
222 hlo_module_config_(hlo_module_config), in DotOpEmitter()
260 /*rhs=*/rhs, /*result=*/target, b_, hlo_module_config_); in EmitTiledLlvmIrGemm()
340 /*result=*/result_op, b_, hlo_module_config_); in EmitTiledLlvmIrGemv()
350 /*result=*/result_op, b_, hlo_module_config_); in EmitTiledLlvmIrGemv()
387 switch (GetDotImplementationStrategy(hlo_module_config_, dot_info_, in Emit()
590 bool multi_threaded = ShouldUseMultiThreadedEigen(hlo_module_config_); in EmitCallToRuntime()
591 bool use_mkl_dnn = hlo_module_config_.debug_options().xla_cpu_use_mkl_dnn(); in EmitCallToRuntime()
Dir_emitter.cc97 hlo_module_config_(hlo_module.config()), in IrEmitter()
101 b_.setFastMathFlags(llvm_ir::GetCpuFastMathFlags(hlo_module_config_)); in IrEmitter()
161 hlo_module_config_, module_, &b_, in InitializeIrFunction()
947 hlo_module_config_, target_machine_features_); in HandleDot()
1192 hlo_module_config_.debug_options().xla_cpu_multi_thread_eigen(); in HandleConvolution()
1194 hlo_module_config_.debug_options().xla_cpu_use_mkl_dnn(); in HandleConvolution()
1286 hlo_module_config_.debug_options().xla_cpu_multi_thread_eigen(); in HandleFft()
1310 if (hlo_module_config_.replica_count() != 1) { in HandleAllReduce()
1881 if (!options::VectorizedReduceDisabled(hlo_module_config_)) { in HandleReduce()
2153 CpuElementalIrEmitter elemental_emitter(hlo_module_config_, this, module_); in HandleFusion()
[all …]
Dir_emitter.h560 const HloModuleConfig& hlo_module_config_; variable
/external/tensorflow/tensorflow/compiler/xla/service/
Delemental_ir_emitter.h39 : b_(b), module_(module), hlo_module_config_(hlo_module_config) {} in ElementalIrEmitter()
211 const HloModuleConfig& hlo_module_config_; variable
Delemental_ir_emitter.cc1601 b_->getInt64(hlo_module_config_.seed() != 0 ? hlo_module_config_.seed() in MakePhiloxRngElementGenerator()
2441 if (hlo_module_config_.replica_count() != 1) { in MakeElementGenerator()
/external/tensorflow/tensorflow/compiler/xla/service/gpu/
Dir_emitter.cc65 hlo_module_config_(hlo_module_config) { in IrEmitter()
76 *hlo, GpuElementalIrEmitter(hlo_module_config_, module_, &b_, in DefaultAction()
158 IrEmitterNested ir_emitter_nested(hlo_module_config_, nested_computation, in EmitCallToNestedComputation()
726 GpuElementalIrEmitter elemental_emitter(hlo_module_config_, module_, &b_, in HandleFusion()
Dir_emitter_unnested.cc590 hlo_module_config_, ir_emitter_context_->llvm_module(), &b_, in HandleFusion()
611 hlo_module_config_, ir_emitter_context_->llvm_module(), &b_, in HandleFusion()
653 GpuElementalIrEmitter elemental_emitter(hlo_module_config_, in HandleFusion()
1008 GpuElementalIrEmitter(hlo_module_config_, module_, &b_, in HandleRng()
1390 VLOG(2) << "AllReduce; replica count: " << hlo_module_config_.replica_count() in HandleAllReduce()
1397 hlo_module_config_.replica_count() > 1 && in HandleAllReduce()
1408 /*replica_count=*/hlo_module_config_.replica_count(), in HandleAllReduce()
1415 if (hlo_module_config_.replica_count() != 1) { in HandleAllReduce()
1420 hlo_module_config_.replica_count(), crs->operand_count(), in HandleAllReduce()
1999 GpuElementalIrEmitter elemental_emitter(hlo_module_config_, in BuildInitializerThunk()
[all …]
Delemental_ir_emitter.h132 const HloModuleConfig& hlo_module_config_; variable
Dir_emitter.h184 const HloModuleConfig& hlo_module_config_; variable
Delemental_ir_emitter.cc74 hlo_module_config_(hlo_module_config), in GpuElementalIrEmitter()
167 if (hlo_module_config_.debug_options().xla_gpu_enable_fast_min_max() && in EmitFloatBinaryOp()