- I . Overview
Hipster is a user-level scheduler for Linux OS, designed to implement a hybrid reinforcement learning algorithm to schedule latency-critical and batch workloads to improve resource utilisation of a cloud system. This scheduler was built at the Barcelona Supercomputing Center and Universitat Politècnica de Catalunya, Spain. The scheduler algorithm should be invoked at the sampling interval of the latency-critical workload. Hipster should be provided with the number of big and small cores, DVFS states of big and small cores, and the static order of the power efficient configurations of the system.
Very important! Please read (at the very least browse through) the following paper before using Hipster. https://nishtalaraj.files.wordpress.com/2016/10/hipster-hpca-camera-ready.pdf The paper will explain the purpose of our scheduler and briefly how it works.
- II. How to install
- Here is how to install packages necessary for Hipster:
sudo pip install psutil
- How to get initial input parameters
To give the basic input details for hipster. Please identify the following:
- The big cores and small cores and their respective DVFS states.
- static_order: Using the stress microbenchmark provided (stress_cpu.c), run it at all possible combination and gather performance (IPS) and power statistics. Only the power-efficiency order (IPS/W) of the combinations for the heuristic mapper.
- MAX_THROUGHPUT, TDP: Run the microbenchmark on all cores at the highest DVFS states and gather performance and power statistics.
- target: target of the latency-critical workload
In case you will have additional questions (especially about internal code of the scheduler), please don't hesitate to email me directly at: [email protected] or [email protected]
- III. Getting help
Got a question? Found a bug? Please contact the me directly.
Please do the following to get a useful response and save time for both of us:
-
Please briefly describe what do you want to use Hipster for? What is the purpose of your experiments? Without understanding what you want to see, it is hard to recommend the best use of Hipster for your task. Also, please elaborate a little bit on the workload you are using in your tests (what apps, what is their CPU usage, etc.).
-
Please indicate what Hipster version you're using, and send the output file from Hipster.
Cheers! -Rajiv