Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I6ba9d0017cc42a4cafe3790e563e206d33c211b3
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|
|
Fixes several issues found by Sonar.
Change-Id: I7ab017f9d89221e8eff4b2438b7669df608b5587
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|
|
Change-Id: I957edd6813b1be43232845d9bf6a4f13ed01ecfb
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|
|
Renames shaded jar name from
jvpp-benchmark-exec.jar
to
jvpp-benchmark-1.18.07-SNAPSHOT-executable.jar
Change-Id: I593946dde3a6fc1717644e7315110aa94553d91d
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|
|
Synchronously creates classify tables.
By default 20x2s warmup
and 100x2s measurement iterations are performed.
VPP is restarted after each iteration.
Each invocation of classifyAddDelTable uses tables
from precomputed set of size tableSetSize.
Tables from the set are used in round-robin fashion.
Run with:
sudo java -jar ./target/jvpp-benchmark-exec.jar \
ClassifyTableCreateBenchmark
To specify aclSetSize (default=100), use:
sudo java -jar ./target/jvpp-benchmark-exec.jar \
ClassifyTableCreateBenchmark -p tableSetSize=1000
To see more options, use
java -jar ./target/jvpp-benchmark-exec.jar -h
Change-Id: I387d879bc99dce45f93d66e8a99f7206f067b443
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|
|
Changes:
- use loopback interface instead of local0,
- if L3 (default) mode is selected, assing IP to the loop0 interface,
- if L2 mode is selected, create bridge domain and assing loop0 the BD.
To test interface in L2 mode, use:
sudo java -jar ./target/jvpp-benchmark-exec.jar -p mode=L2
Change-Id: I8450d42fe0d164c6dbedf1dacf1444a25d9932b8
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
(cherry picked from commit 6e5446c9d64df7f2cbde857d276e8685c9175089)
|
|
Change-Id: I689adb9e139ff22a48cf82e27e8501d542edef72
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|
|
Change-Id: I5ab4465f25c13e0f3b89e2757bde9dd47c660732
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|
|
Creates ACL of size aclSize using acl_add_replace,
then assigns it to local0 using acl_interface_set_acl_list.
Then ACL is updated synchronously using acl_add_replace.
By default 20x2s warmup and 100x2s measurment iterations
are performed.
VPP is restarted after each iteration.
Each invocation of acl_add_replace uses ACL from precomputed
set of ACLs of size aclSetSize.
ACLs from the set are used in round-robin fashion.
Compile:
cd $HC2VPP_ROOT/it/jvpp-benchmark
mvn clean install
Run with:
sudo java -jar ./target/jvpp-benchmark-exec.jar
To specify aclSize (default=100), use:
sudo java -jar ./target/jvpp-benchmark-exec.jar -p aclSize=1000
To specify aclSetSize (default=100), use:
sudo java -jar ./target/jvpp-benchmark-exec.jar -p aclSetSize=1000
To see more options, use
java -jar ./target/jvpp-benchmark-exec.jar -h
Change-Id: I43691ba891940a1c341f8e9893c8598c811bc077
Signed-off-by: Marek Gradzki <mgradzki@cisco.com>
|