CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th ...

10
CSE 598/494 Class 20

Transcript of CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th ...

Page 1: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

CSE 598/494 Class 20

Page 2: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Announcements

Graded midterms handed out Assignment 3 coming up due Nov 14 th

After class collect the hardware platforms Phase 2 reports due Nov 16 th Take home midterm 2 posting date Nov 21

st return date Nov 28 th Final presentations start Nov 28 th

Schedule TBD

Page 3: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Agenda

Assignment 3 discussion Data management continued

Page 4: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Assignment 3

Energy profiling of mobile computing platforms Aim –

Consider the three hardware manifestation of base station Smart phones (Nexus One) Low power hybrid models (Atom based Tunnel Creek) Desktop machine

Power profiling for all the different types of base stations

Page 5: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Computational Workload - BSNBench

BSN applications, some basic tasks are common.

Type of benchmark: Microbenchmark Composition:

Data Operations (Statistics, Differential Encoding) Signal Processing (FFT, Peak detection) Radio Communication (Duty-cycled handshake) Sensor Interface (Sensed Data Query)

Implemented in TinyOS 2.0 and C

Page 6: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Full BSNBench suiteSection Task Example BSN Applications

DATA OPERATIONS

Statistics Glucose Monitoring

Out-of-Range Posture monitoring (accelerometer)

Differential Encoding Temperature Recording

SIGNAL PROCESSING

Fast Fourier Transform Electromyography (EMG) analysis

FIR filtering Motion analysis, De-noising data

Peak detection ECG analysis

RADIO COMMUNICATION

Duty-cycled handshake All wireless BSN applications

Reliable communication Applications with on-body gateway device

SENSOR INTERFACE

Sensed Data Query All sensing applications

Page 7: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Platforms chosen

Personal Laptops Intel Atom based TunnelCreek Platform Nexus One Phones

Page 8: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Assignment Tasks - Coding

Compile and execute the BSNBench tasks in your laptop Can choose mulitple laptops The more the variety of processors the better

Compile and execute the BSNBench tasks in the TunnelCreek platform

Convert the C code for BSNBench to an Android app Run the BSNBench in Google Nexus One phones

Page 9: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Assignment Tasks - Hardware

Power measurement from different platforms For laptops and tunnel creek platforms use a watt meter

For Nexus One Phones Use an ammeter in series

with the battery Measure the power difference

AC Mains

Power Meter

Intel Atom N270 on Mobile Intel Chipset 945 GSE

Page 10: CSE 598/494 Class 20. Announcements Graded midterms handed out Assignment 3 coming up due Nov 14 th  After class collect the hardware platforms Phase.

Submission

Code used for BSNBench in Android Written report with power and energy

consumption results on the different platforms