Search Options

Results per page
Sort
Preferred Languages
Advance

Results 1 - 7 of 7 for gpu_device_metadata (0.26 sec)

  1. tensorflow/compiler/mlir/tensorflow/tests/layout_optimization_layout_assignment_gpu_cc_60.mlir

    // RUN: tf-opt %s -tf-layout-assignment -verify-diagnostics | FileCheck %s --dump-input=always
    
    module attributes {
      tf.devices = {"/device:GPU:0" = #tf_type.gpu_device_metadata<cc_major = 6, cc_minor = 0>}
    } {
    
    // CHECK-LABEL: func @transposeConv2D_3x3_f16
    func.func @transposeConv2D_3x3_f16(%input: tensor<1x28x28x64xf16>, %filter: tensor<3x3x64x64xf16>) -> tensor<1x26x26x64xf16> {
      // cuDNN prefers NCHW data format for spatial convolutions in f16 before
    Registered: Sun Jun 16 05:45:23 UTC 2024
    - Last Modified: Tue Jun 21 08:41:18 UTC 2022
    - 5.8K bytes
    - Viewed (0)
  2. tensorflow/compiler/mlir/tensorflow/tests/layout_optimization_layout_assignment_gpu_cc_70.mlir

    // RUN: tf-opt %s -tf-layout-assignment -verify-diagnostics | FileCheck %s --dump-input=always
    
    module attributes {
      tf.devices = {"/device:GPU:0" = #tf_type.gpu_device_metadata<cc_major = 7, cc_minor = 0>}
    } {
    
    // CHECK-LABEL: func @transposeConv2D_3x3_f32
    func.func @transposeConv2D_3x3_f32(%input: tensor<1x28x28x64xf32>, %filter: tensor<3x3x64x64xf32>) -> tensor<1x26x26x64xf32> {
      // cuDNN prefers NCHW data format for spatial convolutions.
    Registered: Sun Jun 16 05:45:23 UTC 2024
    - Last Modified: Tue Jun 21 08:41:18 UTC 2022
    - 8.5K bytes
    - Viewed (0)
  3. tensorflow/compiler/mlir/tensorflow/ir/tf_structs.h

      void AddGpuDevice(const ParsedName& device,
                        const GpuDeviceMetadata& metadata);
    
      llvm::ArrayRef<ParsedName> device_names() const { return device_names_; }
      size_t NumDevices() const { return device_names_.size(); }
    
      // Returns GPU device metadata if it is available, otherwise returns None.
      std::optional<GpuDeviceMetadata> GetGpuDeviceMetadata(
          const ParsedName& device) const;
    
     private:
    Registered: Sun Jun 16 05:45:23 UTC 2024
    - Last Modified: Thu Mar 02 20:41:19 UTC 2023
    - 2.6K bytes
    - Viewed (0)
  4. tensorflow/compiler/mlir/tensorflow/ir/tf_structs.cc

    }
    
    void RuntimeDevices::AddGpuDevice(const ParsedName& device,
                                      const GpuDeviceMetadata& metadata) {
      device_names_.push_back(device);
      gpu_metadata_.insert({DeviceNameUtils::ParsedNameToString(device), metadata});
    }
    
    std::optional<GpuDeviceMetadata> RuntimeDevices::GetGpuDeviceMetadata(
        const ParsedName& device) const {
    Registered: Sun Jun 16 05:45:23 UTC 2024
    - Last Modified: Thu Mar 02 20:41:19 UTC 2023
    - 1.4K bytes
    - Viewed (0)
  5. tensorflow/compiler/mlir/tensorflow/utils/device_util.cc

      static auto* r = new llvm::Regex("compute capability: ([0-9]+)\\.([0-9]+)");
    
      llvm::SmallVector<llvm::StringRef, 3> cc;
      if (r->match(device.attributes().physical_device_desc(), &cc)) {
        return mlir::TF::GpuDeviceMetadata::get(
            builder->getContext(), std::stoi(cc[1].str()), std::stoi(cc[2].str()));
      }
    
      return builder->getUnitAttr();
    }
    
    // Get devices from an array of string attributes.
    Registered: Sun Jun 16 05:45:23 UTC 2024
    - Last Modified: Thu Apr 25 16:01:03 UTC 2024
    - 6.4K bytes
    - Viewed (0)
  6. tensorflow/compiler/mlir/tensorflow/utils/device_util.h

    // device metadata as a value.
    //
    // Device names added in full parsed device form:
    //   /job:<name>/replica:<replica>/task:<task>/device:<type>:<device_num>
    //
    // Supported device metadata types:
    // (1) GpuDeviceMetadata: GPU device compute capability.
    void AddDevicesToOp(mlir::Operation* op, const DeviceSet* device_set);
    
    // Collects devices information from an op `tf.devices` attributes. Returns
    Registered: Sun Jun 16 05:45:23 UTC 2024
    - Last Modified: Wed Aug 05 20:02:33 UTC 2020
    - 2.3K bytes
    - Viewed (0)
  7. tensorflow/compiler/mlir/tensorflow/utils/device_util_test.cc

      ASSERT_NE(device_meta_0, nullptr);
    
      // GPU device successfully parsed compute capability from description.
      auto device_meta_1 =
          mlir::dyn_cast<mlir::TF::GpuDeviceMetadata>(devices_attr.get(gpu0));
      ASSERT_NE(device_meta_1, nullptr);
      ASSERT_EQ(device_meta_1.getCcMajor(), 7);
      ASSERT_EQ(device_meta_1.getCcMinor(), 0);
    
    Registered: Sun Jun 16 05:45:23 UTC 2024
    - Last Modified: Thu Apr 25 16:01:03 UTC 2024
    - 9.5K bytes
    - Viewed (0)
Back to top