diff options
Diffstat (limited to 'docs/content/release_notes/current/vpp_performance.md')
-rw-r--r-- | docs/content/release_notes/current/vpp_performance.md | 24 |
1 files changed, 5 insertions, 19 deletions
diff --git a/docs/content/release_notes/current/vpp_performance.md b/docs/content/release_notes/current/vpp_performance.md index ea569f547b..56f61ece34 100644 --- a/docs/content/release_notes/current/vpp_performance.md +++ b/docs/content/release_notes/current/vpp_performance.md @@ -3,30 +3,16 @@ title: "VPP Performance" weight: 1 --- -# CSIT 24.06 - VPP Performance +# CSIT 24.10 - VPP Performance 1. TEST FRAMEWORK - - **CSIT test environment** version has been updated to ver. 15, see + - **CSIT test environment** version 15 is used, see [Environment Versioning]({{< ref "../../../infrastructure/fdio_csit_testbed_versioning" >}}). - **General Code Housekeeping**: Ongoing code optimizations and bug fixes. 2. VPP PERFORMANCE TESTS - - Added tests: - - Added memif+DMA tests; added 1518B and 4c memif testcases. - - Added nginx+DMA tests; added 2048B testcases. - - Added IPsec hwasync tests to 3n-icxd and 3n-snr. - - Added IPsec tests to cover more encryption algorithms and other settings. - - Added more SOAK tests. - - Added selected 6-port tests for 3na-spr. - - Edited tests: - - Selected single-flow tests now use single worker even if SMT is on. - - IPsecHW tests now use rxq ratio of 2. - - This means one worker reads only from one of two ports. - - This workaround avoids some inefficiencies, - - but still does not reach the expected performance on 3nb-spr. - - 1518B tests with encapsulation overhead now properly use no-multi-seg. - - Added TX checksum offload to hoststack tests missing it. + 3. PRESENTATION AND ANALYTICS LAYER - - Detailed views added to comparison tables. + - Max value on time axis has been set to utc.now # Known Issues @@ -61,7 +47,7 @@ Issues reported in previous releases which were fixed in this release: # Root Cause Analysis for Regressions -List of RCAs in CSIT 24.06 for VPP performance regressions. +List of RCAs in CSIT 24.10 for VPP performance regressions. Not listing differences caused by known issues (uneven worker load due to randomized RSS or other per-worker issues). Also not listing tests which historically show large performance variance. |