1
0
mirror of https://github.com/pyenv/pyenv.git synced 2025-11-14 06:13:53 -05:00

Updated Home (markdown)

David Dai
2024-01-09 12:26:21 -08:00
parent c88c04b5eb
commit b4d3193e22

38
Home.md

@@ -239,41 +239,3 @@ env PYTHON_CONFIGURE_OPTS='--enable-optimizations --with-lto' PYTHON_CFLAGS='-ma
```
You can also customize the task used for profile guided optimization by setting the `PROFILE_TASK` environment variable, for instance, `PROFILE_TASK='-m test.regrtest --pgo -j0'` will run much faster than the default task.
### Get QPM and AISW
This is needed to run qnn-onnx-converter, qnn-pytorch-converter.
- On local macbook, download QPM [here](https://qpm.qualcomm.com/#/main/tools/details/QPM3). Log in with Qualpass username and password -> select "Linux" -> search QPM.
- `scp` the file to dev machine. For example, `scp QualcommPackageManager.2.0.39.2.Linux-x86.deb weidai@$DEV:/local/mnt/workspace/downloads`
- On Dev Desktop or DevCompute, `sudo dpkg -i <QualcommPackageManager>.deb`
- Install via QPM
```
# log in with Qualpass
qpm-cli --login
qpm-cli --license-activate qualcomm_ai_engine_direct
# This downloads the package
qpm-cli --extract qualcomm_ai_engine_direct
```
If you hit `No usable version of libssl was found`
```
sudo apt-get install libssl-dev
wget http://archive.ubuntu.com/ubuntu/pool/main/o/openssl/libssl1.1_1.1.0g-2ubuntu4_amd64.deb
sudo dpkg -i libssl1.1_1.1.0g-2ubuntu4_amd64.deb
```
In shell (e.g. `~/.bashrc`), add
```
# replace AISW_SDK_ROOT with the version printed from the installation step
export AISW_SDK_ROOT='/opt/qcom/aistack/qnn/2.18.0.240101'
export PATH="$AISW_SDK_ROOT/bin/x86_64-linux-clang:$PATH"
export LD_LIBRARY_PATH="$AISW_SDK_ROOT/lib/x86_64-linux-clang:${LD_LIBRARY_PATH}"
export PYTHONPATH="$AISW_SDK_ROOT/lib/python:$PYTHONPATH"
```