We use our exceptional scientific and engineering background,
interdisciplinary knowledge, strong commitment and motivation,
and unique Collective Knowledge approach
to quickly set up innovative and ambitious R&D projects
while delivering results of highest value, on time.
We help our partners
set up and lead highly innovative and interdisciplinary R&D projects:
- automate performance analysis, benchmarking and optimization of realistic workloads (such as deep learning);
- set up private repositories to crowdsource and reproduce experiments;
- expose all design and optimization knobs from your software and hardware to make it automatically tunable and adaptive;
- simplify optimization knowledge sharing across communities of hardware vendors and software developers;
- perform multi-objective optimization (e.g. cost vs. performance vs. energy vs. accuracy) using practically any programming technologies including OpenCL, OpenMP, CUDA, MPI, and so on;
- apply statistical ("machine learning") techniques (e.g. building performance models, identifying performance bottlenecks);
- develop programming tools (e.g. compilers, profilers, highly optimized and adaptive libraries);
- tune optimization heuristics on representative workloads;
- automatically stress-test compilers;
- optimize predictive models and their topology across realistic inputs;
- perform competitive analysis of SW/HW/model solutions for AI and other emerging workloads.
Live demo of co-designing AI-based applications across the whole SW/HW/model stack using Collective Knowledge:
reusable AI artifacts,
Android app to crowdsource benchmarking of deep learning
across numerous devices from IoT to cloud provided by volunteers similar to SETI@home,
live repository with continouosly aggregated optimization results,
realistic AI training set,
CK-powered compiler crowd-tuning,
ARM demo from TechCon'16.
Do not hesitate to contact us
if you are interested in our services, joint R&D projects and proposals,
or long-term investments to enhance our technology in 2018!