Resource Management in Tessellation OS

1
Energy-Aware Resource Adaptation in
Tessellation OS
Parallel Computing Laboratory
David Chou, Gage Eads Par Lab, CS Division, UC
Berkeley
1. Motivation
5. The Policy Service

• Energy-limited devices execute a mix of
  interactive, high-quality multimedia, and
  parallel throughput applications that compete for
  computing resources
• We want to automatically find low-power,
  high-performance resource configurations while
  providing dynamic adaptation to changing power
  environments

PACORA
Penalty Functions Value of Application to the
System
Runtime Functions Value of Resources to
Applications
Convex Construction

• Two functions represent each application
• penalty function (provided by user)
• runtime function (generated)

2. Basic Goals in Tessellation OS

• Support a simultaneous mix of high-throughput
  parallel, interactive, and real-time applications
• Allow applications to consistently deliver
  performance

Runtime1((0,1), , (n-1,1))
Continuously minimize using the penalty of the
system (subject to restrictions on the total
amount of resources)
3. Space-time Partitioning and Two-level
Scheduling

• A Spatial Partition receives a vector of basic
  resources
• A number of hardware threads, a portion of
  physical memory, a portion of shared cache
  memory, and a fraction of memory bandwidth

6. Energy Control Mechanisms

• A Partition may also receive
• Exclusive access to other resources (e.g., a
  hardware device and raw storage partition)
• Guaranteed fractional services from other
  partitions (e.g., network service)

• Cores
• We can allocate and de-allocate the number of
  hardware threads
• We utilize core idling (C1 state)
• 0 cores idling 170 W
• 29 cores idling 150 W

• Energy Control Knobs
• We have 3 main control mechanisms for
  manipulating the power usage of our evaluation
  system, which is a dual-socket Sandy Bridge
  server with 32 hardware threads.
• Power is measured with a LoCal FitPC
• We also use the Intel on-chip energy counters to
  precisely calculate the power used by the cores
  and DRAM

• Spatial partitioning may vary over time
• Partitions can be time multiplexed resources are
  gang-scheduled
• Partitioning adapts to needs of the system

• 2. Dynamic Voltage/Frequency Scaling
• We have 14 different frequency/voltage states
• At 100 frequency on our evaluation machine, wall
  power measured 250 W
• At 50 frequency, wall power measured 150 W

• 3. Time Multiplexing
• We choose when and how long each cell executes for

• Managing Energy with the Policy Service
• Application 0 can be used to represent the idle
  resources in the system
• Assume all idle resources are powered off
• Alternatively, make penalty a function of the
  power-delay product

?0 is defined to be the total system power ?0 has
a slope that depends on the battery charge

• Scheduling at Level 1 Coarse-grained resource
  allocation and distribution at the cell level
• Scheduling at Level 2 Fine-grained
  application-specific scheduling within a cell

• The Cell Our partitioning abstraction
• User-level software container with guaranteed
  access to resources
• Basic properties of a cell
• Full control over resources it owns when mapped
  to hardware
• One or more address spaces
• Communication channels

4. Resource Allocation Architecture
A Partial and Simplified View

• Cell ID
• Resource counts
• Time parameters

• Distributes resources among cells
• Establishes how cells should be time multiplexed

7. Results
Separation between Mapper and Multiplexer
User Space
Policy Service

• Decision Making Process
• Mapping of multiple resources
• Centralized because it requires global knowledge
• Often expensive

• Initial Adaptation Experiment
• We run two instances of swaptions with initial
  resoures of 16 cores, 16 cpu utilization, and
  maximum CPU frequency. At 25, seconds, we reduce
  frequency to 70 of maximum and cores to 15, as
  proof of concept of power control.
• Power data forthcoming

Space-Time Resource Graph
STRG

• Future Experiments
• Complete and run both energy control mechanisms
  with different applications. Measure system power
  and determine our distance from optimal by
  exploring the resource search space offline.

• Assigns specific resources to cells
• Produces only feasible mappings
• Rejects invalid and infeasible STRGs

Partition Mapping and Multiplexing Layer
Resource Mapper ( STRG Validator)
The Plan

• Execution
• Relatively simple and fast
• If the set of cells does not change
• Allows us to explore decentralized approaches

Tessellation Kernel
Resource Multiplexer

• Determines when cells should be activated and
  suspended
• Actually activates and suspends cells

Partition Mechanism Layer
Research supported by Microsoft (Award 024263)
and Intel (Award 024894) funding and by matching
funding by U.C. Discovery (Award DIG0710227) We
would like to thank the other members of the Par
Lab OS Group, particularly Sarah Bird, Juan
Colmenares, Steven Hofmeyr, and Burton Smith, for
their contributions to the work, as well as the
Akaros team.