Title: Performance Analysis and Monitoring of Volatile
1Performance Analysis and Monitoring of Volatile
Xingfu Wu, Valerie Taylor Department of Computer
Science Texas AM University Third Annual
Workshop on OptIPuter Networing/Backplane
Architecture UCSD, January 25, 2005
2Network Configuration
OPT1 (10.10.1.1)
Optiputer
OPT2 (10.10.1.2)
eth110.10.1.6
eth0 128.194.142.24
Internet cs.tamu.edu
OPT3 (10.10.1.3)
Gigabit Switch (10.0.0.1)
OPT4 (10.10.1.4)
3Snapshot of Volatile on 5-nodes
4Three Components of Volatile
- Volvis main program
- Rendering
- Scalable displaying
- User interaction
- Optistore data server
- Assist visualization dataset handling such as
data management, processing, representation and
transport - Filter raw volumetric data and produce visual
objects - Transfer Function Editor (tfUI)
- User interface for transfer function selection
5Different Access Scenarios
Client1
Volvis
Gigabit Network
Optistore
Client2
Volvis
Master Server
Volvis
tfUI
Optistore
Optistore
Client3
Volvis
Optistore
Client4
Volvis
Optistore
6Different Access Scenarios
Scenario 3
Gigabit Network
Master
tfUI
Volvis
Optistore
7Raw Datasets
8Data access time
9Comparison of three scenarios
10Comparison of three scenarios
11Snapshot of Performance Monitoring
12Short-term Plans
- Port the instrumented Volatile to a large-scale
optiputer testbed for analysis (3/2005) - Analyze the performance of JuxtaView and
LambdaRam applications (6/2005) - Where possible, develop models of data accesses
for the different visualization applications
(9/2005) - Continue collaborating with Jasons group about
viz applications (12/2005)
13Long-term Plans
- Year 3
- Continue with analyzing Volatile on large-scale
OptIPuter testbed - Work with Jason and Luc on redesigning Volatile
- Start working on JuxtaView and LambdaRam
- Year 4
- Work with new Volatile, LambdaRam, JuxtaView on
large-scale OptIPuter testbed - Where possible, develop models of data accesses
for different viz applications - Year 5
- Continue collaborating with Jasons group about
viz applications - Integrate performance monitoring with software
stack to aid in making performance directed
decisions