Nvidia Accelerated Deep Learning
CUDA and cuDNN GPU Acceleration

A comprehensive deep learning suite

Purpose

Complete day-long jobs in 30 minutes. Accelerate Deep Learning up to 20 times (2,000%) the performance of other popular laptops at half the cost. And the Focus has a real-world battery life that can exceed 5 hours.

Please read the disclaimer before proceeding. This workflow is updated regularly. Please write authorship with suggestions or requests.

Overview

Check Libraries (top)

The Focus comes with Nvidia drivers and CUDA drivers tuned to provide excellent deep-learning and gaming performance. However, adding repositories or packages can modify these drivers which can break the deep learning GPU support, Steam and other games, or both. Use the procedure below to check and reset these libraries:

# Run the following in Konsole: /opt/kfocus-001/bin/fixup_run.sh 0 1 1 # If libraries do not match, run the following: /opt/kfocus-001/bin/fixup_run.sh 0 1.1.3-3

Run Conda Script (top)

Use kfocus-001-conda to install an Anaconda suite which is professionally vetted, comprehensive, and GPU accelerated. Open a terminal and enter the following:

kfocus-001-conda

Then follow the prompts. The CUDA 10.0 and 10.1 librares are pre-installed. You will be prompted to consider installing these additional components:

Tests are run at the end of kfocus-001-conda script. This script is non-destructive and can be re-run at any time.

Pycharm is an excellent IDE for Python and may be installed through the Jetbrains toolbox which is pre-installed on all Focus laptops.

Benchmarks (top)

Comprehensive TensorFlow benchmarks can be run at the end of the kfocus-001-conda script. Please compare to the official results.

RTX 2080 RTX 2070 RTX 2060 Titan X [1] Titan XP W10 [1] Titan XP [1]
Interface 8,997 8,295 7,665 7,443 9,741 11,948
Training 9,272 8,824 8,232 7,908 10,375 12,922
AI Score 18,269 17,119 15,897 15,351 20,089 24,870


[1] Desktop GPU

Revisions (top)

This is a minimal revision history. More details are found in git repository.

Disclaimer (top)

We try hard to provide a useful workflow validated by professionals. However, we cannot anticipate every situation, and therefore cannot guarantee this procedure will work for your needs. Always back up your data and test the workflow to determine the correct procedure for you.

THIS WORKFLOW IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS WORKFLOW, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.