Install a comprehensive Python and GPU powered deep learning suite. The Nvidia RTX tensor cores can accelerate deep learning tasks by up to 25x (2,500%) versus other popular developer laptops.
The Focus ships with libraries for deep-learning and gaming. However, adding repositories or packages can modify or break them. You may check and reset these libraries using a terminal as shown below.
Use kfocus-001-conda to install an Anaconda suite. This is professionally vetted, comprehensive, and GPU accelerated. Open a terminal and enter the following.
Follow the prompts. CUDA, TensorRT, and cuDNN libraries are preinstalled.
Pycharm is an excellent IDE for Python. It can be installed through the Jetbrains toolbox which is preinstalled on all Focus laptops.
Tests are run at the end of
kfocus-001-conda. It is non-destructive by default and can be rerun at any time. You may, for example, only install TensorFlow to start and then add PyTorch on a subsequent run. Run the AI suite at the end and compare to the results below.
When running Deep Learning workloads you may notice that the software reports less available RAM than the GPU actually has. Part of this is due to driver overhead, but some memory is used for the actual display. One can use the Intel GPU for display purposes and reserve the Nvidia GPU for Deep Learning or other compute tasks. Start
Nvidia Settings app and then select
Prime Profiles > Nvidia On-Demand and then reboot. This mode does not significantly improve battery life so we recommend using
Nvidia (Performance mode) or
Intel (Power Saving Mode) as it suits your needs at other times.
Comprehensive TensorFlow benchmarks can be run at the end of the
kfocus-001-conda script. Please compare to the official results. RTX 2070 test results current 2020-09-28. Other results are estimates and will be replaced by actual values when testing is complete.
|RTX 2080||RTX 2070||RTX 2060||Titan X||Titan XP W10||Titan XP Ubuntu|
 Kubuntu Focus 2.2Kg, 20mm thick
 Unconstrained Desktop GPU
This is a partial revision history. See the
git repository for all entries.
2020-09-28 5d089610Update to M2 scores
2020-06-10 c4ed9299Restructure layout
2020-05-25 581c6523Expand benchmarks
2020-05-20 d00295d7Add Library check
2020-05-18 8c4b5b82Add GPU library fix
2020-05-15 0684cd5cAdd benchmark
2020-02-18 a4d437c5Add conda script
2020-01-31 70b4aa40Initial document
We try hard to provide a useful solution validated by professionals. However, we cannot anticipate every situation, and therefore cannot guarantee this procedure will work for your needs. Always backup your data and test the solution to determine the correct procedure for you.
THIS SOLUTION 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 SOLUTION, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.