5e4cd1127b
* Added info on DockerHub CI Framework * Feature/azaytsev/change layout (#3295) * Changes according to feedback comments * Replaced @ref's with html links * Fixed links, added a title page for installing from repos and images, fixed formatting issues * Added links * minor fix * Added DL Streamer to the list of components installed by default * Link fixes * Link fixes * ovms doc fix (#2988) * added OpenVINO Model Server * ovms doc fixes Co-authored-by: Trawinski, Dariusz <dariusz.trawinski@intel.com> * Updated openvino_docs.xml * Edits to MO Per findings spreadsheet * macOS changes per issue spreadsheet * Fixes from review spreadsheet Mostly IE_DG fixes * Consistency changes * Make doc fixes from last round of review * integrate changes from baychub/master * Update Intro.md * Update Cutting_Model.md * Update Cutting_Model.md * Fixed link to Customize_Model_Optimizer.md Co-authored-by: Trawinski, Dariusz <dariusz.trawinski@intel.com> Co-authored-by: baychub <cbay@yahoo.com>
2.2 KiB
Using GPU Kernels Tuning
GPU Kernels Tuning allows you to tune models, so the heavy computational layers are configured to fit better into hardware, which the tuning was done on. It is required to achieve best performance on GPU.
Note
Currently only convolution and fully connected layers undergo tuning process. It means that the performance boost depends on the amount of that layers in the model.
OpenVINO™ releases include the <INSTALL_DIR>/inference_engine/bin/intel64/Release/cache.json file with pretuned data for current state of the art models. It is highly recommended to do the
tuning for new kind of models, hardwares or drivers.
Tuned data
GPU tuning data is saved in JSON format. The file is composed of 2 types of attributes and 1 type of value:
- Execution units number (attribute): splits the content into different EU sections
- Hash (attribute): hashed tuned kernel data
- Key (value): Array with kernel name and kernel's mode index
Usage
You can activate Kernels Tuning process by setting KEY_TUNING_MODE flag to TUNING_CREATE and KEY_TUNING_FILE to <"filename"> in a configuration map that is
passed to the plugin while loading a network.
This configuration modifies the behavior of the ExecutableNetwork object. Instead of standard network compilation, it will run the tuning process.
Please keep in mind that the tuning can be very time consuming. The bigger the network, the longer it will take.
File with tuned data is the result of this step.
Note
If a filename passed to
KEY_TUNING_FILEpoints to existing tuned data and you are tuning a new model, then this file will be extended by new data. This allows you to extend existingcache.jsonprovided in the OpenVINO™ release package.
The example below shows how to set and use the key files:
@snippet snippets/GPU_Kernels_Tuning.cpp part0
You can activate the inference with tuned data by setting KEY_TUNING_MODE flag to TUNING_USE_EXISTING and
KEY_TUNING_FILE flag to <"filename">.
GPU backend will process the content of the file during network compilation to configure the OpenCL kernels for the best performance.