Future TableView features

[FloVanGH]

trafficstars

This is relevant after #2032 is finished:

• [x] Resizing of columns
• [ ] Build in column sort mechanism (e.g. by internal generate sort model wrapper)
• [ ] More cell types like selection (compare ComboBox), and bool (CheckBox)
• [x] Optional row zebra pattern (show/hide by flag)
• [x] Draw native table rows with QStyle.

[tilakrayal]

@vishwascm, The difference in code execution time between a TensorFlow installation from source and a pre-built binary installed using pip install tensorflow can be influenced by several factors.

When you are trying to build TensorFlow from source, there is an option to customize compiler flags and optimizations. If the compilation is not configured optimally for your hardware, it might result in performance. Also the speed of the storage device where TensorFlow is installed can affect the loading and execution times. If the source build is on a slower storage medium compared to the system where the pre-built binary is installed, it can contribute to differences in execution time.

To improve the execution time for TensorFlow built from source, consider the following:

- Use appropriate compiler flags for your CPU architecture.
- Check and customize the build configuration to match your requirements.
- Ensure that GPU support is configured correctly if applicable.
- Monitor the build logs for any errors or warnings that might indicate issues.

TensorFlow build from source can sometimes lead to increased code execution time compared to using pip install tensorflow. While building from source offers certain advantages like customization and access to the new features, it can also introduce performance drawbacks if not done correctly. Thank you!

[vishwascm]

Hi @tilakrayal, Is there any place where I can see the build command along with all configuration used and flags used for official tf v2.15.0 wheel file available in pypi.org for aarch64 machine (graviton) with sve_256?

[vishwascm]

Hi @tilakrayal , Is there any place where I can see the build command along with all configuration used and flags used for official tf v2.15.0 wheel file available in pypi.org for aarch64 machine (graviton) with sve_256?

[tilakrayal]

@vishwascm, Apologies for the delay. You can directly install the tensorflow v2.15 on aarch64 using pip install tensorflow-aarch64 where the flags are not required. https://pypi.org/project/tensorflow-aarch64/

If you are trying to install the tensorflow using a build source, then we will use the flags based upon the requirement. https://www.tensorflow.org/install/source#build_the_package

Thank you!

[github-actions[bot]]

This issue is stale because it has been open for 7 days with no activity. It will be closed if no further activity occurs. Thank you.

[vishwascm]

Hi @tilakrayal, Thanks for the reply. I am looking for some more detail about building tensorflow from source, specific to aarch64 architecture with vector length 256 (Graviton) machine, including clang version, ubuntu version, build tags to use aarch64 etc. Just like how official wheel files are built. Are these details available in tensorflow git repository itself? If available how to use it.

Thanks, Vishwas

[tilakrayal]

@vishwascm, Could you please take a look at this official build from the source document where you can find the clang version, flags required to install the tensorflow. https://www.tensorflow.org/install/source

sudo apt-get update && sudo apt-get install -y llvm-16 clang-16

You can try to specify the flags for aarch64 as per the requirement to install the tensorflow from build.

Thank you!

[github-actions[bot]]

This issue is stale because it has been open for 7 days with no activity. It will be closed if no further activity occurs. Thank you.

[github-actions[bot]]

This issue was closed because it has been inactive for 7 days since being marked as stale. Please reopen if you'd like to work on this further.

[google-ml-butler[bot]]

Are you satisfied with the resolution of your issue? Yes No