Home
TuneUp.ai Performance Tuning-as-a-Service
Artificial Intelligence-based, Fully Automatic Parameter Tuning Solutions for Servers, Data Centers, Mobile Devices, IoT, and Supercomputers
Most computers are not at their best performance
Slow websites or apps that cannot handle requests smoothly
Expensive electricity and cooling bills
Whole system brought down by small workload spikes
Why is performance tuning hard?
Traditional performance tuning usually costs more than what you gain
- One computer can have hundreds of tunable parameters. A complex cluster could have thousands.
- Default parameter values are for easy setup, not performance.
- Must be done by high-paid domain experts.
- Tuning time and cost comprise a big part of your service bill.
- You can’t know the tuning effect before spending money to do it.
TuneUp.ai Brings the Latest AI-based Tuning to You in Three Steps
1. Sign up at tuneup.ai — Register an account on tuneup.ai
2. Download the client agent — Download and install our lightweight client agent
3. Done! — Watch the performance of your system go higher in one or two days
Versatility of Our Products
Applies to many scenarios
You pick the tuning-up goal
We charge by how much you use
Our Technologies
Deep Reinforcement Learning has beaten the best human players in games like StarCraft and Go
Our Deep Reinforcement Learning-based tuner is capable of testing and measuring all parameter combinations on its own in a highly efficient manner
We do all our complex machine learning tasks online, relieving customers from having to understand those onerous tasks
TuneUp.ai LightningCloud vs. Manual Tuning
| TuneUp.ai LightningCloud | Manual Tuning | |
|---|---|---|
| Skill requirement | Regular IT staff | Vendor and domain experts |
| Total cost | Low and predictable | High and unpredictable |
| Applicable time | 24x7x365, no downtime needed | Require planned downtime |
| Required time | 10 hours to a few days during regular operation | Days or weeks downtime |
| Tuning effect | Most workloads will see improved performance | Depends on the skills of the experts |
| Handling rapidly changing workloads | Real time | Depends on response speed of staff |
| Safety | High | Depends on the process and staff |
| Handling hardware or software change | Automatically | Need to redo the tuning |
Contact Us
Contact us if you wish to know more about our product roadmap and release plans. Thanks!
Also see the Pilot Benchmark Framework and the TuneUp.ai White Paper.