SLIDE 1
1
Testing Proof on Xen
Rosa M Garcia Sverre Jarp
2
Index
- Proof communication
- Different scenarios
- Results on PC with processes
- Results on PC with domains (Xen)
- Comparison
- Conclusions
Testing Proof on Xen Rosa M Garcia Sverre Jarp 1 Index Proof - - PowerPoint PPT Presentation
Testing Proof on Xen Rosa M Garcia Sverre Jarp 1 Index Proof - - PowerPoint PPT Presentation
Testing Proof on Xen Rosa M Garcia Sverre Jarp 1 Index Proof communication Different scenarios Results on PC with processes Results on PC with domains (Xen) Comparison Conclusions 2 Master Proof communication
SLIDE 2
SLIDE 3
3
Slave … Initialize Process Process Process Wait for Next command Master Packet Generator Add Histograms Display Histograms Slave … Initialize Process Process Process Wait for Next command
Tree->Draw
GetNextPacket
Tree->Draw
GetNextPacket GetNextPacket GetNextPacket GetNextPacket GetNextPacket SendObjectHisto SendObjectHisto
Proof communication
SLIDE 4
4
Different test scenarios
- 4 slaves + master distributed in 2 Pc running as processes
- 4 slaves + master running in independent domains
Fedora C4 Xen SLave PROOF + Master SLave PROOF Fedora C4 Xen Slave PROOF SLave PROOF Guest domain Fedora SLave PROOF + Master SLave PROOF Fedora SLave PROOF SLave PROOF Process
SLIDE 5
5
Benchmark
- The benchmark provides 2 selectors, each
- ne reading a different amount of data
– ProcOpt Reads 25% of the data. – Proc Reads all the data
- Measurements in the Slave
– Packet processing time – Get Packet Latency – CPU Time
SLIDE 6
6
First results in scenario 1
Benchmark: Reads only 25% of the data Scenario: 2 PC with each slave is running in a process The graphics include error bars per value
SLIDE 7
7
First results in scenario 1 (CPU Time)
Good values Small error
SLIDE 8
8
Second Benckmark in scenario 1
Benchmark: Reads all the data Scenario: 2 PC with each slave is running as a process
Good values Small error
SLIDE 9
9
How things should be
Why those huge errors?
Range 0.2 -2 Error 0.355
SLIDE 10
10
How things are
Why those huge errors?
Range 0.2-22.5 Error 3.385
Range0.2-16.6 Error 2.462
Range -28.3 Error 4.063
Range0.2-55 Error 7.844
SLIDE 11
11
First benchmark in scenario 2(Xen)
Benchmark: Reads only 25%
- f the data
Scenario: 2 PC with each slave runs on a guest domain
SLIDE 12
12
Second Benchmark in scenario 2
SLIDE 13
13
Graphical comparison Xen-PCs
SLIDE 14
14
Graphical comparison Xen-PCs
SLIDE 15
15
Comparison
0.415625 0.69751 0.0444994
Proc
0.386286 0.810273 0.0341873
Opt
0.771708 0.771708 0.00664648
Proc
0.418625 0.767716 0.0060823
Opt CPU Proc Latency
XEN PROCESS
Unreliable comparison because of the errors
SLIDE 16
16
Conclusions
- As the error in the data is significant the
result (CPU time) is not relevant – no realistic information.
- In the communication , last transmitions
are enormous, and introduce a lot of noice
– Bottleneck in the master – Possible solution move the master to another node
SLIDE 17
17