e61c9b10cb868e6148b6e7271b856e948241705d |
|
03-Oct-2017 |
Todd Kjos <tkjos@google.com> |
Fix off-by-one error with binderThroughputTest -s The -s option to binderThroughputTest allows the user to set the size of the payload being sent during the test. Due to an off-by-one error, the size was actually 4 bytes less than specified. Fixed. Test: verified correct size Bug: 65052700 Change-Id: I884fe08c113c2ca63a6f328741fb0914799e9eb7 Signed-off-by: Todd Kjos <tkjos@google.com>
/frameworks/native/libs/binder/tests/binderThroughputTest.cpp
|
3b17a3128f1bc044eadee17b546b29394ef791e3 |
|
03-Oct-2017 |
Todd Kjos <tkjos@google.com> |
Fix binderThroughputTest -m option The -m option allows the user to specify the estimated max latency to calibrate buckets used for percentile calculations. The option didn't work because the time_per_bucket was not being set so the default value was used. Test: verified correct -m operation Bug: 65052581 Change-Id: I96d6ba0271edaffda0bd9cccfef5d559d3bf111c Signed-off-by: Todd Kjos <tkjos@google.com>
/frameworks/native/libs/binder/tests/binderThroughputTest.cpp
|
c88f8bb54a39902206e7353dcd1a12bb5f07ddb2 |
|
15-Jun-2017 |
Sherry Yang <sherryy@google.com> |
Fix percentile calculations Adjust max latency in running binderThroughputTest to improve percentile accuracy. We used to use a fixed max latency which is devided into time buckets to obtain the round trip latency percentile. The actual latency, however, can be much smaller than the max latency, which results in measurements from all iterations ending up in the same bucket and leads to inaccurate percentile measurement. This is resolved by either performing a training round (option -t), which sets the max latency to two times the worst latency during training, or by providing a max latency (option -m) in millisecond by the caller. Use --help option to see usage. Bug: 62660944 Change-Id: Iaa23837eb24b90012ae20c35f20caae33986b35a Fixes: 62660944 Test: Running "binderThroughputTest -p -w 2 -i 10000 -s 0" on marlin-3.18 used to show the same percentile values for 50%, 90%, 95%, and 99%. Now with -t flag,percentile values are different. Signed-off-by: Sherry Yang <sherryy@google.com>
/frameworks/native/libs/binder/tests/binderThroughputTest.cpp
|
25e1ded18ca51db209f3b8c3e199b4945066ab0c |
|
23-Nov-2016 |
Todd Kjos <tkjos@google.com> |
Add options to specify payload size and use client/server pairs Test: tested manually Change-Id: I744d84b1621c2dd19525d3a65ed543d9dea575d4
/frameworks/native/libs/binder/tests/binderThroughputTest.cpp
|
515aa0b911aebb0bd4665bc0b61a693d6951e994 |
|
11-Aug-2015 |
Riley Andrews <riandrews@google.com> |
binder: Add a throughput performance test for binder. Change-Id: I04801cd505a439307f2a963f6267db1831b3829f
/frameworks/native/libs/binder/tests/binderThroughputTest.cpp
|