21.4% regression in vol25_09_loopback_32bit in build s10u9_09
vol25_09_loopback_32bit :

--------------------------------------------------------------------------------
Historical vol25_09_loopback_32bit results on oaf408_64
--------------------------------------------------------------------------------
Build Result Status Description Related Bugs
--------------------------------------------------------------------------------
s10u9_09 87475.08 RED 21.4% regression
s10u9_08 111353.8 GREEN 27.8% improvement
s10u9_07
s10u9_06
s10u9_05 87138.05 RED 23.3% regression
s10u9_04
s10u9_02
s10u9_01
s10u8_08a 113651 GREEN unchanged
s10u8_08
s10u8_07 113743 GREEN unchanged
s10u8_06
s10u8_05 113655 GREEN 1.3% improvement
s10u8_04
s10u8_03 112197.6 RED 1.3% regression 6862074
s10u8_02 113644.5 RED 3.6% regression 6858973
s10u8_01 117869.8 GREEN 3.1% improvement
s10u7_08 ~ YELLOW Comparison not statistically significant
s10u7_07 ~ YELLOW Comparison not statistically significant
s10u7_06 ~ YELLOW Comparison not statistically significant
s10u7_05 ~ YELLOW Comparison not statistically significant
s10u7_04 ~ YELLOW Comparison not statistically significant
s10u7_03 ~ YELLOW Comparison not statistically significant
s10u7_02 113851.8 GREEN 1.3% improvement
s10u7_01
s10u6_07b 112371.4 First Result

Using s10u9_08 as baseline
8 baselines. Average is 111353 PctStDevis 1.26%
Plot_mean: 87475.08 , overall change: 21.44 %
Iteration Result Change Colour
1 87375 -21.53 % RED
2 88117 -20.87 % RED
3 87457 -21.46 % RED
4 86708 -22.13 % RED
5 87667 -21.27 % RED
6 87523 -21.40 % RED
Number of runs: 6
With all numbers shown, overall change will be -21.44%
Previous build:s10u9_08 111353 0.00 % YELLOW
The found results are: 1 YELLOW, 6 RED including the old result
Std. Deviation as % from Mean is 0.52
TTest against baseline: 0.0000000000

Outliers
=========
1 low outliers in test
1 high outliers in test
Can't find the machine in /etc/bootparams.
Benchmark res_range: 10:-1 (not ship)
Driver Version changes...
  -mac MAC Services v1.10 (1.10 -> 1.1)
  -dld Data-Link Driver v1.6 (1.6 -> 1.7)
  -dls Data-Link Services v1.7 (1.7 -> 1.8)
  -elfexec exec module for elf 1.121 (1.121 -> 1.122)
  -fctl SunFC Transport v20100331-1.59 (1.59 -> 1.60)
  -qlc SunFC Qlogic FCA v20090929-2.32 (2.32 -> 3.00)
  -mac_ether Ethernet MAC plugin 1.1 (1.1 -> 1.2)
  -sd SCSI Disk Driver 1.556 (1.556 -> 1.557)


Detailed statistical comparison of results is available at:
http://perfwww.ireland/cgi-bin/detail.cgi?rig=oaf408_64&benchmark=vol25_09_loopback_32bit&build=s10u9_09&base=s10u9_08&clean=FALSE

Using s10_74L2a as baseline
No Baselines
Plot_mean: 87475.08 , overall change: 100.00 %
Warning: plot_per is 0.00, but overall change is 100.00. Will update later.
Iteration Result Change Colour
1 87375
2 88117
3 87457
4 86708
5 87667
6 87523
Number of runs: 6
With all numbers shown, overall change will be 0.00%
Previous build:s10u9_08 111353
Std. Deviation as % from Mean is 0.52
TTest against baseline: 1.0000000000

Outliers
=========
1 low outliers in test
1 high outliers in test
Can't find the machine in /etc/bootparams.
Benchmark res_range: 10:-1 (not ship)

Detailed statistical comparison of results is available at:
http://perfwww.ireland/cgi-bin/detail.cgi?rig=oaf408_64&benchmark=vol25_09_loopback_32bit&build=s10u9_09&base=s10_74L2a&clean=FALSE

http://perfwww.ireland/cgi-bin/publictools/shipII.cgi?&BUILD=s10u9&BNAMES=vol25_09_loopback_32bit&tSysteM=oaf408_64

The machine is a :X5355 8 x 2.66GHz 4 GB ram
Looking at the results history here. They look dependant on the Java version.

Build Version Result
s10u8_03 java version "1.5.0_19" 112416
s10u8_05 java version "1.5.0_20" 113089
s10u8_07 java version "1.5.0_20" 111798
s10u8_08a java version "1.5.0_20" 113819
s10u9_05 java version "1.6.0_18" 86620.4
s10u9_08 java version "1.5.0_24" 112180
s10u9_09 java version "1.6.0_19" 87375.9
s10u9_10 java version "1.6.0_19" 88702
s10u9_12 java version "1.6.0_21" 103355
s10u9_13 java version "1.6.0_21" 103904
There was a perf jump here from
1.6.0_19 to 1.6.0_20 but comparing 1.5.0_24 to 1.6.0_21
we are still 6% down.
I'll reduce the size / priority of this CR and confirm
if this regression is still present as of s10u9_14a
I have confirmation that the 1.6.0 JDK on s10u9_14a is still
performing 6% worse than the 1.5.0 JDK bundled with the same build.


Using s10u9_14a-6954933-jdk1.5.0_24 as baseline
12 baselines. Average is 110562 PctStDevis 1.20%
Plot_mean: 103977.71 , overall change: 5.96 %
Iteration Result Change Colour
1 102252 -7.52 % RED
10 104372 -5.60 % RED
11 103868 -6.06 % RED
12 102958 -6.88 % RED
13 103981 -5.95 % RED
14 103626 -6.27 % RED
15 103760 -6.15 % RED
16 104548 -5.44 % RED
17 104489 -5.49 % RED
2 104204 -5.75 % RED
3 103971 -5.96 % RED
4 104371 -5.60 % RED
5 103721 -6.19 % RED
6 103704 -6.20 % RED
7 104135 -5.81 % RED
8 105184 -4.87 % RED
9 104477 -5.50 % RED
Number of runs: 17
With all numbers shown, overall change will be -5.96%
Previous build:s10u9_13 104011 -5.93 % RED
The found results are: 18 RED including the old result
Std. Deviation as % from Mean is 0.64
TTest against baseline: 0.0000000001

Outliers
=========
1 low outliers in test
Can't find the machine in /etc/bootparams.
Benchmark res_range: 10:-1 (not ship)

Detailed statistical comparison of results is available at:
http://perfwww.ireland/cgi-bin/detail.cgi?rig=oaf408_64&benchmark=vol25_09_loopback_32bit&build=s10u9_14a&base=s10u9_14a-6954933-jdk1.5.0_24&clean=FALSE