Complete day-long jobs in 20 minutes. Accelerate TensorFlow deep learning up to 25 times (2,500%) faster than other popular developer laptops that cost twice a much.
Please read the disclaimer before proceeding. This workflow is updated regularly. Please write authorship with suggestions or requests.
The Focus ships with libraries for deep-learning and gaming. However, adding repositories or packages can modify break them. You may check and reset these libraries using a terminal as shown below:
kfocus-001-condaUse 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 10.0 and 10.1 libraries are pre-installed. Prompt will pop up asking you to consider installing these additional components:
(GPU)cuDNN libraries (required)(GPU)TensorRT libraries (required)(GPU) MXNet(GPU) PyTorch(GPU) TensorFlow(GPU) Theano(GPU) KerasPycharm is an excellent IDE for Python. It can be installed through the Jetbrains toolbox which is pre-installed on all Focus laptops.
Tests are run at the end of the script. 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.
Comprehensive TensorFlow benchmarks can be run at the end of the kfocus-001-conda script. Please compare to the official results.
| RTX 2080[1] | RTX 2070[1] | RTX 2060[1] | Titan X[2] | Titan XPW10[2] | Titan XP[2] | |
|---|---|---|---|---|---|---|
| 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] Kubuntu Focus 2.2Kg, 20mm thick
[2] Unconstrained Desktop GPU
Q: Starting Nvidia Visual Profiler results in an error. How can I fix this?
A: One must install Java 8 and then call nvvp with with correct path.
This is a partial revision history. See the git repository for all entries.
2020-06-10 c4ed9299 Restructure layout2020-05-25 581c6523 Expand benchmarks2020-05-20 d00295d7 Add Library check2020-05-18 8c4b5b82 Add GPU library fix2020-05-15 0684cd5c Add benchmark2020-02-18 a4d437c5 Add conda script2020-01-31 70b4aa40 Initial documentWe 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.